This application is a continuation-in-part of United States patent application Serial No. 105,926, filed October 7, 1987.

Technical Field The invention relates to a method of diagnosing and monitoring immunologic disease activity such as AIDS-related complex (ARC) and acquired immunodeficiency syndrome (AIDS), and to the identification and monitoring of a subset of CD4 ⁺ cells which is preferentially infected with immunodeficiency viruses, using a class of monoclonal antibodies directed against organ-specific leucocyte adhesive receptors for endothelium (or HEV-type receptors).

Background of the Invention Most normal lymphocytes recirculate continuously throughout the lymphoid organs of the body. The recirculation is nonrandom, that is, the lymphocytes exit from the bloodstream into lymph nodes, Peyer's patches, and certain sites of chronic inflammation, such as arthritic synovia, in an organ- specific manner. Such exiting is mediated by a specific adhesive interaction between the blood-borne lymphocyte and the lining of the postcapillary high endothelial venules (HEV) at these sites. This adhesion is organ-specific in character and is mediated by functionally distinct receptors on the lymphocyte surface which recognize ligands on the targeted vascular endothelium and facilitate attachment of the lymphocyte. For a review, see Jalkanen, S., et al. Immunological Reviews 91:39-60, 1986, hereby incorporated by reference.

At least three distinct specificities of receptors have been described on cloned lymphoma populations that exhibit unispecific HEV binding potential. That is, some lymphomas bind only to Peyer's patch HEV, some bind only to lymph node HEV, and others bind exclusively to venules found in arthritic synovia. The organ specificity of these lymphoma populations was exploited by using them as immunogens, in order to derive the first monoclonal antibody,

MEL-14, that defines the lymph node-specific "homing receptor" found on murine lymphocytes and lymphoma cells. This work began in 1981 in the laboratory of Dr. Irving Wiseman at Stanford University in collaboration with Dr. Eugene Butcher. See Gallatin, W. M., et al. Nature 303:30, 1983, hereby incorporated by reference. Subsequent publications have described the distribution of this antigen during normal and neoplastic development of the immune system in mice. See Gallatin, W.M., et al., Cell 44:673-680, 1986, hereby incorporated by reference.

The MEL-14 antibody recognizes a 90 kd molecular weight glycoprotein, expressed on murine lymphocytes and lymphoma cells, which is composed of approximately 40 kd worth of N-Iinked carbohydrate but apparently no O-linked carbohydrate and a polypeptide core of approximately 45 to 47 kd molecular weight. This moiety also appears to be covalently modified by the addition of the small highly conserved polypeptide ubϊquitin. See Seigelman, et al., Science 231:823, 1986, and St. John, et al., Science 231:845, 1986, both of which are hereby incorporated by reference.

Dr. Judy Woodruff and her eoworkers have recently described in the rat system other monoclonal antibodies that recognize similar, although not absolutely identical in terms of structure, receptors that appear to mediate similar functions in that system. For a review, see Woodruff and Clarke, Ann. Rev. Immunol. 5:201, 1987, hereby incorporated by reference. To date, no sequence information exists which would allow direct comparison of the core peptides of the different organ specificities, but it seems likely at this point that the structures are probably members of a multigene family of closely-related receptor species distinct from those typified by the CD11/CD18 series of cell adhesion structures. For a review, see Springer, et al., Ann. Rev. Immunol. 5:223, 1987, hereby incorporated by reference.

Based on what was known about the organ distribution, cell line distribution, and molecular weight of the MEL-14 antigen (gp90 ^MEL ~ ¹⁴ ), another series of monoclonal antibodies was prepared in Dr. Eugene Butcher's lab at Stanford which define, in humans and in the macaque, cell surface glycoproteins that are structurally and functionally homologous to gp90 ^MEL . The prototype antibody of this series, Hermes-l, differs from MEL-14 qualitatively in that it recognizes a framework epitope that is found on a class of 95 kd molecular weight glycoproteins present on human and primate lymphocytes, as well as other hemopoietic cells. [gp95 Hermes-l is a distinct molecular species from the B chain of the FA type cell adhesion molecules (which are also known by WHO

nomenclature as CD-18).] That is to say, Hermes-l recognizes a series of closely- related cell surface glycoproteins, one member of which appears to be functionally homologous to gp90 MEL-14 ^ _{e >} j _t πιe<_iϊa.tes attachment to lymph node-type HEV). Also, while the antibody MEL-14 blocks the attachment of murine lymphocytes to lymph node HEV, Hermes-l, by virtue of the type of epitope it recognizes, does not block the attachment of lymphocytes to high endothelial venules. However, second generation antibodies raised against purified Hermes-l antigen derived from human cells will block the attachment of lymphocytes to high endothelial venules (personal communication from Eugene Butcher).

Comparative analysis of the Hermes-l antigens with those recognized by anti-extracellular matrix type III receptor monoclonal antibodies indicates a strong structural homology between these two classes of molecules. A cDNA encoding one form of the Hermes-l antigen has recently been isolated. Mouse lymphocyte cells transf ected with full length versions of this clone express a glycoprotein of 85-90 kd which reacts with Hermes-l, anti-ECMK III monoclonal antibodies and with monoclonal antibody A1E3, an anti-CD4 specific antibody. Therefore, the glycoproteins recognized by these reagents can be encoded by a single gene; i.e., they are structurally related or identical. A structural link between the murine and human systems has thus been described. MEL-14, the antibody, does cross-react weakly with the human Hermes-l antigen and, in fact, the MEL-14 antibody will block the attachment of human lymphocytes to high endothelial venules under appropriate conditions.

Immunologic disorders involving the immune system can be manifested by a variety of disease states, including congenital immunodeficiency, multiple sclerosis, rheumatoid arthritis, and systemic lupus erythematosus. See, e.g., Chapter 2, The Merck Manual, 14th Ed., 1982, hereby incorporated by reference. In particular, immunodeficiency diseases encompass a diverse group of conditions, characterized chiefly by an increased susceptibility to various infections with consequent severe acute, recurrent, and chronic disease, which typically result from deficiency of one or more components of the immune system, e.g., B cell or T cell or other leucocyte imbalances. Methods are being explored to measure the host's immune response in such disease states. Such immunological monitoring includes enumeration of lymphocyte subpopulations; see, e.g., U.S. Patent No. 4,677,061, entitled T-Cell Lymphocyte Subset

Monitoring of Immunologic Disease.

Acquired immunodeficiency syndrome (AIDS) is caused by the Human Immunodeficiency Viruses (HIV), a group of lymphotropic lentiviruses. Two distinct classes of such viruses have been isolated from patients with AIDS. Viruses isolated from patients in the U.S., Europe, and Central Africa have been designated HIV-1 (formerly HTLV-III, LAV, and ARV). More recently, viruses isolated from West African patients have been identified as HIV-2 (formerly LAV-2).

The different HIV-1 isolates share the same biological properties as well as antigenically cross-reactive proteins. Some nucleotide sequence variation exists between North American and Central African isolates and to a lesser degree between independent isolates in the U.S. HIV-2 is antigenically quite distinct from HIV-1. There is some cross-reactivity between the gag (core protein) and pol (polymerase) gene products of the two virus groups, but little if any cross-antigenicity among the env (envelope) products. Nucleotide sequence data has confirmed that although they are genetically related, HIV-1 and HIV-2 are quite distinct. About 58% to about 59.4% of the amino acids of HIV-2 gag and pol, respectively, are identical to the corresponding HIV-1 polypeptides.

Simian immunodeficiency viruses (SIVs) related to these human AIDS viruses have been isolated from nonhuman primates. The first of these, isolated from rhesus monkeys, were termed STLV-lll _mae , and subsequently other isolates have been obtained from asymptomatic African green monkeys (STLV- IIIg _m ) and mangabey monkeys (SIV _mm ). The University of Washington Regional Primate research center isolate is also a lentivirus, termed SIV _mne , having been isolated from the human HUT-78 cell line after cocultivation with lymphocytes from a pigtailed macaque, Mαcαcα nemestrina, that died of a malignant lymphoma in 1982. Benveniste, R.E., et al., J. Virol. 60:483-490, 1986, hereby incorporated by reference. The various SIV isolates are antigenically related to one another. Comparison of the restriction maps and gag region nucleotide and amino acid sequences of SIV _mne and SIV _m reveals that these isolates, thought not identical, are approximately 94% homologous (Benveniste, personal communication). The genetic relationship between the SIV and HIV families of AIDS viruses is only starting to be assessed, but it is clear that the two groups are very closely related. Nucleotide sequence analyses reveal that HIV-2 is more closely related to SIV than to HIV-1. The gag, pol, and env proteins of SIV and HIV-2 are antigenically cross-reactive, but SIV/HIV-2 cross-reactivity to HIV-1 is restricted to certain gag and pol epitopes.

Perhaps the single most striking feature of both HIV and SIV is their tropism for CD4+ cells. The CD4 antigen is an important component of the cellular receptor for HIV, and the major virus envelope glycoprotein, gpl20, binds directly to the CD4 molecule. Virus binding involves only some CD4 epitopes and the virus' binding site for CD4 appears to be relatively conserved between different HIV isolates. The affinity of gpl20 for CD4 plays an integral role in the in the formation of syncytia by infected lymphocytes, leading to cell death of these pivotal components of the body's immune system.

In normal macaques, CD4+ cells constitute about 30-55% of blood lymphocytes. However, in situ hybridization analyses reveal that only a very small fraction of lymphocytes express viral RNA at any given time. This suggests that certain CD4+ lymphocytes (or stages of their differentiation) may be particularly susceptible to infeetion/replication. In humans, virus has been found primarily in CD4+ cells expressing the Leu8 marker, but this isn't that useful since Leu8 is expressed on 85-90% of CD4+ cells. In other studies, Giorgi et al. (J. Immunol. 138(ll):3725-3730, 1987) failed to find evidence for a selective loss of cells bearing either the CDw29, CD45R, or Hb-11 markers.

The genome of the virus used in the studies described below, SIV _mne , has been cloned. A comparison of the gag gene sequence of SIV _mne with those of HIV-1 and HIV-2 indicates an 86% homology to HIV-2 and only a 45% homology to HIV-1 (Benveniste, personal communication). SIV is thus very closely related to HIV-2 which causes an AIDS-like illness in humans. The genomic: similarity of SIV and HIV extends to their biological behavior. SIV _mne like other SIVs, causes a fatal, immunosuppressive disease: experimentally infected animals suffer from profound anemia, skin rash, generalized wasting, prolonged fever, Iymphadenopathy, nephrosis, opportunistic infections, and show a selective decrease in CD4+ lymphocytes in blood before death (Benveniste, et al., J. Virol., in press, 1987). HIV-1, which does not infect macaques, will replicate in chimpanzees albeit with a nonpathogenie outcome compared to humans. With the severe shortage of chimpanzees, that model has limited utility. Given the close relationship in genomic structure and pathobiology of SIV and HIV-2, infection of macaques with SIV isolates, such as SIV _mne , is currently the best available animal model for human AIDS.

Summary of the Invention The invention provides a method of diagnosing and monitoring disease activity in a mammalian host. Leucocytes from the host are contacted

with two types of antibodies, both of which are directed against leucocyte differentiation antigens (e.g., CD1-CD45 in WHO nomenclature), at least one of which is a leucocyte cell surface receptor for vascular endothelium. The ratio of the first and second antigens on the host cells is determined, e.g., with a flow cyto meter using fluorophore-labeled antibodies, and the ratio is compared to a healthy control or host-specific standard as indicative of disease activity. In a representative embodiment, the invention provides immunological binding partners directed against a cell sur ace receptor for vascular endothelium characteristic of human CD4+ T cells that are lost preferentially during HIV infection. More specifically, the subject immunological binding partner is capable of resolving a pool of circulating human CD4+ T cells into a first population of CD4+ cells preferentially lost during HIV infection which are more susceptible to HIV infection and a second population of CD4+ cells which are deleted mueh later or not at all in the course of HIV infection and which are intrinsically less susceptible HIV infection or cytolysis. The immunological binding partner, which may be an antibody, antibody binding fragment, etc., bears an immunoglobulin antigen- binding site conferring the requisite specificity and is typically provided with a detectable marker, such as a fluorophore or radionuclide. Kits are also provided for carrying out the subject method. Description of the Drawings

FIGURE 1 is a contour plot of a two-color flow cyto metric analysis (gated to exclude monocytes and granulocytic cells by forward versus 90° light scatter) of normal macaque peripheral blood lymphocytes (PBL), wherein Hermes- 1 is plotted on the y-axis and CD4+ is plotted on the x-axis, illustrating the division of CD4+ cells into Hermes-l" ¹ and Hermes-l ⁰ subpopulatϊons;

FIGURE 2 is a set of three histograms of one-color flow cytometric analyses of the distribution of Hermes-l on normal macaque lymphocytes in blood (PBL), thoracic duct lymph (TDL), and mesenteric lymph node (MNL), showing that Hermes-l cells predominate over Hermes-l cells in blood but that the reverse relationship occurs in lymph;

FIGURE 3 is a line graph illustrating the mitogenie response (measured via the incorporation of H-thymidine into newly synthesized DNA) of CD4+/Hermes-l ^hl vs. CD4+/Hermes-l ^l0 lymphocytes, isolated via fluorescence activated cell sorting (FACS) and subjected to varying doses of the phorbol ester, TPA, demonstrating that the Hermes-l ^hl cells responded approximately ten times more vigorously as a population than the Hermes-l cells;

FIGURE 4 is a FACS contour plot of normal macaque PBL showing Hermes-l affinity (y-axis) versus forward light scatter (x-axis), indicating that the Hermes- l ^hl subset contains most of the larger and putatively dividing cells;

FIGURE 5 is a pair of single-parameter histograms of Hermes-l FITC-labeled whole PBL from healthy (A) and clinically ill SIV-inf ected (B)

Macaque nemistrina primates, wherein cell number (y-axis) is plotted versus

Hermes-l (x-axis), illustrating the hi to lo Hermes-l shift that characterizes the onset and course of SIV/HIV infection;

FIGURE 6 is a comparison of two-color FACS plots of PBL (excluding monocytes and granulocytes) from a healthy control Macaca fasicularis (top) and an SIV/Mne-infected M. fasicularis (bottom), wherein the y-axis indicates staining with a phycoerythrin-conjugated anti-CD4 reagent and the x-axis indicates staining with an FITC-labeled Hermes-l reagent, demonstrating that Hermes-l cells are absent from the CD4+ pool in the infected animal but present in the control animal;

FIGURE 7 presents equivalent analyses to those in FIGURE 6 but here for CD4 and CD45 in SAIDS-D-infected animals; and

FIGURE 8 is a tabulation of absolute percentages of total PBL as well as Hermes-l ^hl ' ¹⁰ ratios in CD4+ lymphocytes in control, SIV-infected, and SAIDS-D-infected macaques, illustrating how loss of the Hermes-l ^nι /CD4+ cells characteristically occurs prior to catastrophic loss of CD4+ cells in SIV-induced AIDS yet is not an obligatory correlate of poor health status in infection via SAIDS-D, an unrelated D type retrovirus.

Detailed Description of the Preferred Embodiment The invention serendipitously resulted from investigations of the cross-reactivity of the monoclonal antibody Hermes-l with lymphocytes from various primate species, particularly the pigtailed macaque, Macaca nemistrina, and the crab-eating macaque, M. fasicularis, that were undertaken to develop this primate model for study of lymphocyte recirculation in a system closely related to man. Thus, the distribution of the Hermes-l marker versus other available lymphocyte differentiation markers (CD1, CD3, CD4, CD8, etc.) was assayed on macaque lymphocytes obtained from different organ sources. During the course of these investigations, two discrete levels of expression of the Hermes-l marker were surprisingly observed on normal macaque T lymphocytes. These two levels of expression, which are termed "hi" and "lo" throughout this disclosure, are especially distinct in the T-lymphocyte subsets

defined by the CD4 and CD8 markers. FIGURE 1 is a representative data plot that illustrates the characteristic division of CD4+ normal macaque peripheral blood lymphocytes into Hermes-l" ¹ and Hermes-1 ° subpopulations. At present, the precise role of these two levels of expression, in terms of the in vivo migratory behavior of these cells, is unknown. But, characteristically, their distribution within distinct anatomic compartments of the immune system differs. As shown in FIGURE 2, in peripheral blood, Hermes-l ^hl cells predominate in normal macaques, while in thoracic duct lymph and mesenteric lymph node Hermes-l ¹⁰ cells predominate. Within lymph nodes, the distinction between Hermes-l and Hermes-l cells is less precise, and there is some blending of the two phenotypes.

Two discoveries concerning the expression of the Hermes-l antigen on macaque T-cells indicate that the HEV marker recognized by this type of antibody is useful in diagnosing and predicting the clinical outcome of infection with AIDS-associated lentiviruses such as human immunodeficiency virus/simian immunodeficiency virus (HIV/SIV). First, as shown in FIGURE 3, CD4+/Hermes- l" ¹ lymphocytes and CD4+/Hermes-l ^lc> lymphocytes differ markedly (by > 10X) in their prolϊf erative response to mitogenϊc stimuli. As a population, the cells of the Hermes-l" ¹ subset are much more readily activated to a higher level of mitogenic response by phorbol esters, such as TPA, and other agonistic signals. Shown here are the prolif erative states of the two subpopulations, assayed via incorporation of a radiolabeled DNA precursor at four days poststimulation. FIGURE 4 shows that the Hermes-l subset of macaque peripheral blood lymphocytes contains most of the larger and presumptively dividing cells. The intrinsic background proliferation in culture of CD4+/Hermes-1" ¹ cells is also slightly higher.

Second, in a series of macaques experimentally infected with SIV/Mne (at the University of Washington Regional Primate Research Center), animals which became clinically ill and eventually died showed a selective loss of Hermes-l" ¹ lymphocytes in their bloodstream. This characteristic shift from preponderance of the hi to lo subsets is illustrated in FIGURES 5, 6, and 8 and may reflect either a redistribution of the Hermes-l" ¹ cells into other organs, a conversion of the Hermes-l cells to the Hermes-l phenotype, and/or selective cytolysis of the Hermes-l" ¹ cells, presumably as a result of the viral infection. Regardless of the explanation, this characteristic shift is correlated with a clinical outcome of SIV infection. The following points can be made regarding these findings: (1) loss of Hermes-l cells is not simply a marker for viremia per

se; (2) loss of Hermes-l" ¹ cells from the CD4+ pool precedes the catastrophic loss of CD4+ cells in some SIV-infected individuals; and (3) loss of Hermes-l cells is not inexorably linked to immunosuppression or poor health since it does not occur in animals infected with an unrelated d-type virus, SAIDS-D (FIGURES 5, 7, and 8). The loss of Hermes-l ^hl cells may also occur in CD8+ cells and, perhaps, even the B cell subsets.

It is clear that the disease caused by SIV in the macaque is quite similar if not identical in its histopathology to that observed in humans following HIV-l/HIV-2 infection. In both syndromes there is a selective deletion of CD4+ - cells from the blood, and virus is exclusively recovered from CD4+ cells in the macaque system. Given the fact that isolation of infectious virus in vitro from lymphocytes appears to require lymphocyte activation, and that Hermes-l and Hermes-l ° CD4+ cells in the macaque differ in their response to activating agents, it seems likely that the predelection of Hermes-l cells to proliferate in vivo might predispose them to cytolytic infection via SIV and explain their disappearance in vivo as a class.

These hypotheses were directly tested by separately isolating the

Hermes-l and Hermes-l CD4+ cells from infected animals and assaying directly for the presence of SIV in these two subpopulations. Similarly, the hi and lo cells were isolated from uninfected individuals and tested for differential capacity to mount a productive infection in vitro. Evidence for productive infection was observed exclusively in the Hermes-l fraction of cells isolated from: infected animals. The homologous population obtained from control animals also showed the same differential susceptibility to infection after exposure to virus ύτ vitro. The same subsets obtained from infected animals were tested for presence of SIV related DNA sequences by amplification of a portion of the GAG region of SIV/Mne using polymerase catalyzed thermal recycling (PCR). In this case, the oligonucleotide primers represented sequences conserved between SIV,

HIV-1 and HIV-2. Even though at the time of assay Hermes-l ^hl cells represented 1-2x10 of the total peripheral blood mononuclear cells, all the detectable SIV sequences were localized to this subset. Ultimately, based upon the foregoing seminal observations, the cellular, genetic and biochemical differences which define resistance/susceptibility to infection in these two populations of cells may be defined and manipulated to modify consequences of viral infection. For example, subtraetive or differential cDNA libraries representing the steady state transcriptional differences between Hermes-l and Hermes-l cells of the same lineage can be readily constructed and compared.

For the present purposes, application of Hermes-1-type antibody reagents for the diagnosis and prognosis of clinical AIDS/ARC is reasonably straightforward. First, Hermes-1-type reagents can be useful in identifying the subset of CD4+ cells in humans that is most relevant to monitor for the isolation of HIV and for charting the progression of the disease or the detection of virus. Normally, only about one cell in 10 ⁴ reads out as an infected cell in typical assays that are currently performed to monitor the presence of HIV per se in AIDS patients. Such assays are typically performed using whole peripheral blood lymphocytes: the frequency of infected cells within the Hermes-type-hi CD4+ subset is significantly higher than this. That is to say, it is contemplated that the signal-to-noise in assays for virus would be enhanced significantly if one were to examine the most relevant target population for the virus. Second, since the observed loss of Hermes-l" ¹ CD4+ cells in the macaque system corresponds to progression of SIV infection, a capacity to monitor the presence of cells having this phenotype in the blood is of value in predicting the clinical outcome of HIV infection in humans and as a means for monitoring the efficiency of therapeutic regimens. Third, there are therapeutic benefits to be derived from selectively removing the HIV-infected CD4+ lymphocytes from the .bloodstream of AIDS patients. Thus, the subject antibody reagents can be employed in available immunoselective devices and systems, e.g., bonded to an insoluble substratum within a chromatography column for on-line or off-line filtration of the buffy coat layer from plasma, to preferentially remove CD4+ cells susceptible to HIV infection from a patient's bloodstream while preferentially retaining CD4+ cells substantially less susceptible to HIV_ infection. Unfortunately, the subdivision of T-cell subsets via the Hermes-l antibody is somewhat species-specific and can be accomplished in macaques, but not always reliably with human cells using this particular reagent. This probably reflects a phenomenon related to the type of epitope the Hermes-l antibody recognizes in human versus nonhuman primate systems, rather than any fundamental difference in the T cells that are present in the human and macaque systems. A large body of evidence suggests that virtually all of the CD series lymphocyte differentiation markers presently defined in human are also present in macaque and define similar, if not identical, types of lymphocytes. See Clark et al., Immunogenetics 18:599, 1983. Thus, although the Hermes-l antibody itself may not be the optimum antibody for this purpose in humans, because it does not always allow similar division of human CD4+ cells, given the above observations in

-li¬

the macaque, second generation monoclonal antibodies that will allow such a distinction in human cells can be readily produced and identified.

For example, hybridomas producing monoclonal antibodies directed against a cell surface receptor for HEV characteristic of human CD4+ T cells that are preferentially depleted from the bloodstream during HIV infection are raised as follows.

Mice are immunized with tonsillar lymphocytes of human or

Q macaque origin by standard techniques. For example, two immunizations (2 x 10 cells) can be made in Freund's adjuvant or phosphate buffered saline (PBS), two to fifteen days apart. Spleen cells are fused three days later with available myeloma cell lines, such as SP20 or NS1, again using standard techniques such as polyethylene glycol, and the fusion products are plated out in 96-well plates.

Alternatively, the immunization can be made with a Hermes-1- type antigen isolated by immunoaffinity chromatography using the Hermes-l or other similar antibody (e.g., ECMR III or CD44 antibodies). For example, lymphocytes are first lysed in 3% NP40 detergent Tris-NaCl and centrifuged 10 min at 1500 x g. The supernatant is collected and centrifuged 30 min at 30 k. The resulting supernatant is precleared . overnight with albumin-sepharose and then subjected to affinity selection with WGA (wheat germ agglutinin)-sepharose. Bound material is eluted with 200 mM N-acetyl-d-glucosamine and then affinity selected with Hermes- 1-type antibody-sepharose. Elution can be made with either 0.5 M propionic acid, 25 mM octyl-glucoside, pH 2.8, or 0.2 M sodium carbonate, 25 mM octyl-glucoside, pH 11. Sodium duodecyl sulfate (SDS) is added to 0.1%, and the purified antigen is dialized against Tris buffer and lyophilized. The antigen is employed in conventional immunization and hybridoma production protocols, such as those described above with the HEV antigen-bearing cells. Here the mice can be immunized with the immunoselected antigen either retained on the antibody-coated Sepharose beads or in pure form as a slice out of a polyacrylamide gel. The resulting hybrids can be screened in the following manner. As a first tier screen, hybridoma supernatants can be tested for reactivity with either the purified antigen (by radioimmunoassay or Western blot) or by immunofluorescence or radioimmunoassay on the cell used (e.g., tonsillar lymphocytes, or any human or primate Hermes-1-reactive cell line) for preparation of the original immunizing antigen. Hybridomas positive in the first tier selection are then subjected to second tier screening which can involve

two-color flow cytometric analysis of human peripheral blood lymphocytes, with the reactive hybridoma supernatants in one color (fluorochrome) versus antibodies directed against CD4 (and possibly CD8) coupled to a different fluorochrome. One simply looks for hybrids which divide the CD4+ cells into two populations with a frequency and magnitude of signal (i.e., 1:1 to 3:1 and particularly on the order of 2:1 in favor of the more highly labeled subpopulation) similar to that shown in FIGURE 1 with Hermes-l and CD4 in the macaque system. Hybrids passing this test are subjected to third tier analysis, which can take one of two forms.

In the first case, one isolates the "new hybrid hi" CD4+ cells and "new hybrid lo" CD4+ cells identified above via standard cell sorting techniques and determines whether the hi subpopulation responds more readily to mitogenic signals provided by agents such as TPA. Alternatively, one can add a Hoechst viable dye to the above analysis with CD4 and the new hybrid for three-color flow cytometric studies. From work in the macaque system, we know that the Hermes- l ^nι subset contains most, if not all, of the actively cycling CD4+ cells found in peripheral blood. The Hoechst dye allows one to establish position within the eell cycle, so its addition to this screening permits one to identify hybrids in which the "new hybrid hi" CD4+ cells contain the bulk of the dividing cell population within the helper cell subset. Some suggestion that this is the case would be obtained as matter of course during the second tier screening since a shift in forward scatter (i.e., cell size) would be observed in simple two-color analyses. In either case, differential mϊtogen response activity of the hi and lo subpopulations on the order of that shown in FIGURE 3 (i.e., >10X hi to lo with TPA) will confirm the suitability of the antibody. Additional confirmation may be made by using the selected antibody to radioimmunoprecipitate cell surface iodinated material from human lymphocytes; the species recognized is around 90 to around 100 kD molecular weight, and typically on the order of 95 kD. Alternatively, the r elatedness of the molecular species recognized by any new hybrid with that recognized by Hermes-l could be obtained in a straightforward fashion by conducting radioimmunoprecipitations wherein the new hybrid is tested for its capacity to preclear the Hermes-l defined antigen from lysates of Hermes-l positive cells.

While the invention has been described above in terms of an exemplary embodiment, it is to be understood that the subject method can be used to diagnose and monitor a wide variety of immunologic disorders in addition to AIDS. Diagnosis can thus be made of autoimmune disorders of a systemic, e.g.,

rheumatoid arthritis, or organ-specific nature, e.g., insulin-dependent diabetes, as well as intermediate conditions such as pernicious anemia. Hypersensitivity reactions can also be diagnosed and monitored by the disclosed method, as can congenital (primary), acquired, as well as secondary immunodeficiency diseases. Post-transplantation events can also be systemically monitored, e.g., for rejection of transplanted kidney, heart, liver, or bone marrow and/or engraftment of bone marrow. Additionally, the therapeutic response of a patient to various chemotherapeutic, radiotherapeutic and other clinical treatments can be monitored by the subject method; for example, the immune response of a cancer patient to chemotherapy can be thereby monitored for optimization of the patient-specific treatment regimen.

Similarly, various techniques may be employed for determining the ratios of cells having the specified pattern of antigen expression. A wide variety of techniques exist for measuring the presence of specific antigens on cells using a wide variety of detectable markers or labels, such as radionuclides, fluorescers, chemilu inescers, particles, enzymes, enzyme substrates or cofactors or inhibitors, paramagnetic metals, or the like. At the present time, for determination of the presence of a multiplicity of epitopic sites on a specific cell, the preferred technique is multiparameter flow cytometric analysis (Parks, Lanier and Herzenberg, Flow Cytometry and Fluorescence Activated Cell Sorting (FACS), in: A Handbook of Experimental Immunology, 4th Ed., Weir, D. M., Chap. 29, Blackwell Scientific Publications, London). This method involves conjugating, either directly or indirectly, antibodies specific for different epitopic sites with fluorescers, different fluorescers being used for each site, and employing a cell sorter with multicolor analysis. By employing fluorescers having different emission peaks, one provides for independent detection of each of the epitopic sites.

Of particular interest is the employment of fluorescers providing for long Stokes shifts (>25nm), absorbing below about 520 nm, preferably below about 500 nm and emitting above about 525 nm. The antibodies to the specific surface membrane proteins may be directly conjugated to the fluorescers, or they may be indirectly labeled, that is, the fluorescer may be covalently linked to the. antibody or noncovalently linked through ligand-receptor complex formation, e.g. hapten-antibody, anti-Ig antibody, or biotin-avidin. For example, the antibody can be biotinylated and reacted with fluorescer-conjugated avidin.

Various fluorescers find use, such as fluorescein, rhodamine, Texas red, phycobiliproteϊns, such as phycoerythrin, allophycocyanin, phycocyanin, phycoerythrocyanin, and the like, umbellif erone, dansyl, etc.

In carrying out the assay, blood samples are taken from a human or other mammalian patient, particularly peripheral blood, and the desired cells isolated by conventional techniques and suspended in an appropriate medium. Particularly, the blood sample may be introduced into a heparinized receptacle, diluted 1:1-1.5 in a conventional tissue culture medium, layered on a lymphocyte separation medium such as Ficoll-Hypaque, and the interface mononuclear cells washed and resuspended at a concentration of about 10 _ to 108 /ml in a tissue culture medium with appropriate adjuvants. Alternatively, a whole blood sample in which the red blood cells have been lysed by addition of an appropriate agent, such as ammonium chloride, can be utilized. The cell suspension may then be incubated at moderate temperatures (-5° to 25° C) with the appropriate antibodies (in two stages where the labeling is indirect), these cells washed and then freed of agglomerated cells, conveniently by passing through a micropore filter. The amount of antibody employed will typically be about 1-3 doubling dilutions above the titratϊon end point. The suspension is now ready for use in a cell sorter or analyzer in accordance with conventional techniques. For diagnosing the probability of a change in disease activity, one can determine the mean value of the normal population and choose a range of one or more standard deviations from the mean to provide for greater certainty as to the absenee of false positives and false negatives. Conveniently, from one to three standard deviations, preferably about two standard deviations is employed from the arithmetic mean, so that the normal range covers four standard deviations from the mean (±2SD). The statistics may be further refined by providing for further subdivisions to determine normal values, such as dividing the patient groups by age, sex and other statistically significant criteria. Values outside of the normal range, either high or low, are indicative of a change in disease activity, while values restoring the ratio to the normal range are indicative of a remission to homeostasis. Thus, by monitoring changes in the indicated ratios, one can predict with reasonable probability the onset of clinical illness or a remission and relate this to the appropriate therapeutic regimen.

The immunologic binding partners which find use are antibodies, particularly monoclonal antibodies, as well as FAB fragments and the like, specific for one or more epitopes of surface membrane proteins as specified

herein. For example, where one of the monoclonal antibodies that are employed provides for distinguishing the CD4+ subset of T-cells, such antibodies may be obtained by immunizing an appropriate host, conveniently a mouse, with, e.g., a HPB-ALL T cell line (human peripheral blood - acute lymphocytic leukemia) and screening for antibodies specific for detecting the CD4 helper/inducer T cell- associated antigen. Such monoclonal antibodies against leucocyte differentiation antigens are also commercially available from a variety of sources.

The other antibodies to the membrane surface proteins associated with cell homing and adhesion may be obtained by immunization of an appropriate host as described above.

To prepare monoclonal antibodies, the host will normally be given booster shots, the host's spleen isolated, and fusions carried out in accordance with conventional techniques. (See, for example, Kennett et al., Monoclonal Antibodies, Plenum Press, NY, 1980, and the references cited therein.) The antibodies may be any of the immunoglobulin types, for the most part they will be IgG, K of λ, usually <, and may be IgGl, 2a, 2b, or 3, of murine or other origin.

Kits can be provided for detection of the cell populations to determine the previously indicated ratios. The kits may include the antibodies for one or both aspects of the ratio and, depending upon the particular protocol, may have the antibodies labeled or unlabeled. Where unlabeled, each of the antibodies will usually be from a different host, so that labeled anti-immunoglobulin may be employed, which will bind to only one of the antibodies to allow for detection of the presence of the particular epitopic site. Alternatively, the antibodies may be conjugated with different ligands, e.g. biotin, in which case fluorescer-conjugated receptors, e.g. fluorescing avidin, are employed.

The monoclonal antibodies may be provided in a single composition, conveniently lyophilized and combined with appropriate additives, such as stabilizers, photobleach retardants, buffers, e.g. Tris, phosphate, etc., where the amount of antibody will be reconstituted prior to use to provide the desired concentration of antibody. Usually, the number of different fluorescers which will be present in a single mixture will be not greater than about six, more usually not greater than about four, generally ranging from two to four, preferably from two to three. Thus, where the ratio is dependent upon the analysis of the presence of two epitopic sites, one can provide one or two mixtures of antibodies depending upon the ability to distinguish between the fluorescence of the different

antibodies. Fluorescent combinations of particular interest include fluorescein and phycoerythrin, phycoerythrin and allophycoeyanin, fluorescein and Texas Red, etc. Other materials which may be included with the antibodies or in combination in the kit include lymphocyte separation medium, photobleach retardants, and various washing buffers, such as phosphate-buffered saline optionally containing bovine serum albumin.

While the preferred embodiments of the invention have been illustrated and described, it is to be understood that, within the scope of the appended claims, various changes can be made therein. Hence, the invention can be practiced in ways other than those specifically described herein.