Method for screening anti-ligand libraries for identifying anti-ligands specific for differentially and infrequently expressed ligands

Abstract

The present invention relates to screening methods and, in particular, to methods of screening anti-ligand libraries for identifying anti-ligands specific for differentially and/or infrequently expressed ligands. The method comprises the steps of providing a library of anti-ligands: providing a first subtractor ligand; providing a second target ligand; determining the amount of the first and second target ligands using one or more equations derived from the universal law of mass action; providing the determined amount of a first subtractor ligand; providing the determined amount of a second target ligand; providing separation means capable of use to isolate anti-ligand bound to the second target ligand from anti-ligand bound to the first subtractor ligand; exposing the library of to the first and second target ligands to permit binding of anti-ligands to ligands; and using the separation means to isolate the anti-ligand bound to second target ligand.

Claims

1. A method of isolating an anti-ligand specific for a differentially expressed target ligand, said method comprising the steps of: (1) exposing a library of anti-ligands to an amount of a subtractor ligand construct and an amount of a target ligand construct, to permit binding of said anti-ligands specific for a differentially expressed target ligand, (i) wherein said subtractor ligand construct comprises a subtractor ligand commonly expressed on said subtractor ligand construct and on said target ligand construct with equal density, (ii) wherein said target ligand construct comprises said subtractor ligand, and further comprises said target ligand having unique or upregulated expression on said target ligand construct compared to said subtractor ligand construct (iii) wherein said amount of the subtractor ligand construct and said amount of the target ligand construct are determined using the following equation, or derivative thereof: ${\mathrm{bA}}_p=\left\{\frac{\left(A+T+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}{2}-\sqrt{\frac{{\left(A+T+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}^2}{4}-\mathrm{AxT}}\right\}x\left\{\frac{\left(T_p{\mathrm{xC}}_p\right)}{\left(\left(T_p{\mathrm{xC}}_p\right)+\left(T_s{\mathrm{xC}}_s\right)\right)}\right\}$ where bA.sub.p=The number of bound anti-ligand to said target ligand construct at equilibrium T.sub.p=The number of target ligand per target ligand construct T.sub.S=The number of target ligand per subtractor ligand constructs C.sub.p=The number of target ligand constructs C.sub.s=The number of subtractor ligand constructs A=Total number of anti-ligand T=Total number of ligands of both target ligand constructs and subtractor ligand constructs C=Avogadro's constant (6.02210 particles/mole) V=Reaction volume in liters Kd=Equilibrium dissociation constant for target ligand; and (2) isolating the anti-ligand bound to the target ligand construct, thereby isolating said anti-ligand specific for said differentially expressed target ligand.

2. The method of claim 1 wherein said amount of the subtractor ligand construct and said amount of the target ligand construct are determined automatically, or wherein step (1) or (2) are performed automatically.

3. The method of claim 1 further comprising the step of separating the anti-ligand bound to the ligand of the target ligand construct.

4. The method of claim 1 wherein steps (1) and (2) are repeated one or more times.

5. The method of claim 1 wherein the amount of the subtractor ligand construct is provided between 10 and 1000 fold in excess of the amount of the target ligand construct.

6. The method of claim 5 wherein the excess of the amount of the subtract ligand construct is 10 to 100 fold.

7. The method of claim 1 wherein said subtractor ligand construct and said target ligand construct are independently selected from the group consisting of a solid support, cell membrane and/or portions thereof, synthetic membrane, beads, and chemical tags.

8. The method of claim 7 wherein said subtractor ligand construct or said target ligand construct is cell membranes and/or portions thereof.

9. The method of claim 8 whereby ligands on the subtractor ligand construct and ligands on the target ligand constructs are fixed to and/or incorporated within separate cell membranes and/or portions thereof.

10. The method of claim 1 whereby the subtractor ligand construct and the target ligand construct have a different density.

11. The method of claim 10 wherein the subtractor ligand construct is of a lower density than the target ligand construct.

12. The method of claim 11 wherein the subtractor ligand construct is a membrane vesicle.

13. The method of claim 11 wherein the target ligand construct is a whole cell membrane.

14. The method of claim 1 wherein step (2) is performed using at least one method selected from the group consisting of: density centrifugation, solid support sequestration, magnetic bead sequestration, chemical tag binding, and aqueous phase partitioning.

15. The method of claim 14 wherein step (2) is performed by density centrifugation.

16. The method of claim 15 wherein the density centrifugation utilizes a sucrose-polymer gradient.

17. The method of claim 1 wherein the library of step (1) is a display library comprising a plurality of library members which display anti-ligands.

18. The method of claim 17 wherein the library is a phage display library.

19. The method of claim 1 wherein the ligand on the subtractor ligand construct and the ligand on the target ligand construct are independently selected from the group consisting of: (i) antigens; (ii) receptor ligands; and (iii) enzyme targets selected from the group consisting of: a carbohydrate; protein; peptide; lipid; polynucleotide; inorganic molecules; and conjugated molecules.

20. The method of claim 1 wherein the library of anti-ligands comprises one or more of the following: antibodies and/or antigen binding variants, derivatives or fragments thereof; scaffold molecules with engineered variable surfaces; receptors; and/or enzymes.

21. The method of claim 1 comprising a further step of exposing the ligand of the target ligand construct to a stimulus which influences the expression of said ligand on said target ligand construct.

Description

DESCRIPTION OF PREFERRED EMBODIMENTS AND DRAWINGS

(1) Examples embodying certain aspects of the invention shall now be described, with reference to the following figures in which:

(2) FIG. 1The capture of anti-ligands complexed to ligand as a function of target ligand density where: Kd=10 nM, target ligand specific anti-ligand copy number (A)=200, volume=2.5 ml.

(3) FIGS. 2A and 2BThe anti-ligand molecule capture on target ligand constructs as a function of 1) the number of ligands expressed on the subtractor and target ligand constructs respectively and 2) the number of target and subtractor ligand constructs used in the selection process. The indicated numbers represent the capture of anti-ligands (on target ligand constructs) with specificity for ligands represented in Table 1. The reaction parameters Kd=10 nM, target ligand specific anti-ligand copy number T=200 for all anti-ligand specificities, volume (V)=2.5 ml.

(4) FIG. 3Anti-ligand capture on target cells as a function of anti-ligand specificity and affinity. Antigen categories are as described in Table 1. Target cell input (C.sub.P) is 510.sup.7, subtractor cell input (C.sub.S) is 10.sup.10, volume (V) is 2.5 ml and assuming a ligand specific anti-ligand input (A) of 200.

(5) FIGS. 4A and 4BHistogram showing frequencies of anti-ligand specificity as a function of excess subtractor ligand (cells) added and the number of selection rounds performed. Histograms show the relative numbers of anti-ligands isolated with different specificities. The reaction parameters of FIGS. 2A and 2B apply.

(6) FIG. 5Target population ligand abundance independent, but target to subtractor ligand population ligand abundance dependent isolation of anti-ligands with specificity for Iigands of different abundance within target ligand population and between target and subtractor ligand populations as described in Table 2. The reaction parameters of FIG. 2 apply, and subtractor ligand population excess is 200 as described in example 4.

(7) FIG. 6Elution of CD40L specific phage and general mouse fibroblast binder phage following selection on non-transfected mouse fibroblast cells (3T3), CD32 transfected mouse fibroblast cells (CD32-3T3) or CD40L transfected mouse fibroblast cells (CD40L-3T3). Dotted lines show the theoretical number of CD40L specific phage bound to CD40L-3T3 cells at equilibrium given a CD40L specific phage particle input number of 10.sup.5 or 10.sup.6 and parameters used in the selection process.

(8) FIGS. 7A & 7B Numbers (FIG. 7A) and ratios (FIG. 7B) of CD40L specific phage particles to general fibroblast binder phage (determined by cfu after infection of E. coli HB101F and plating on LA plates containing appropriate antibiotics), following selection in absence of subtractor ligands (CD40L-3T3 cells only), pre-selection using a fourfold excess of CD32-transfected 3T3 cells, and pre-selection using a 50-fold excess of cell membranes from CD32 transfected 3T3 cells.

(9) FIGS. 8A & 8B B cell (Ramos) versus T cell (Jurkat) reactivity of scFv clones: (FIG. 8A) Hit rate of clones selected following three rounds of competition bio-panning utilising the invention (FIG. 8B) Hit rate of clones selected following three rounds of conventional negative (T cell) pre-selection and following positive (B cell) selection. Y axis=B cell reactivity of clones (counts*fluorescence intensity). X axis=T cell reactivity of clones (counts*fluorescence intensity). ScFv binding was examined using an FMAT macro-confocal high throughput screening instrument.

EXAMPLE 1

Deriving Equations

(10) Applying the Universal Law of Mass Action (LMA), the number of ligands needed to isolate anti-ligands to low expression ligands and/or differentially expressed ligands from display libraries of high diversity may be calculated.

(11) The LMA states that the non-covalent (hydrogen bonding, electrostatic, Van der Waals or hydrophobic forces), reversible binding between an anti-ligand A and its target ligand T, and their complex AT is given by the equilibrium interaction A+T AT with the equilibrium dissociation constant or affinity K.sub.d=[A][T]/[AT].

(12) The equilibrium interaction between anti-ligands with identical specificity (A) for a target ligand (T) may be described as
Bound A (bA)free A (fA)+free T (fF)
with

(13) $\begin{array}{cc}\mathrm{Kd}=\left[\mathrm{fA}\right]x\frac{\left[\mathrm{fT}\right]}{\left[\mathrm{bA}\right]}& \left(I\right)\end{array}$

(14) It is known that the total A or T is the sum of free and bound A or T i.e. [A](Total A)=[fA]+[bA], and [T](Total T)=[fT]+[bA] Therefore in (I) replacing [fA] by [A][bA], and [fT] by [T][bA]

(15) $\begin{array}{cc}{K}_d=\left(\left[A\right]-\left[\mathrm{bA}\right]\right)x\frac{\left[T\right]-\left[\mathrm{bA}\right]}{\left[\mathrm{bA}\right]}& \left(\mathrm{II}\right)\end{array}$
Which is rearranged to form
(K.sub.d[bA])=([A][T][A][bA])([T][bA][bA].sup.2)
0=[bA].sup.2([A]+[T]+K.sub.d)[bA]+[A][T]
This simultaneous equation has the solution

(16) $\left[\mathrm{bA}\right]=\frac{\left(\left[A\right]+\left[T\right]+K_d\right)}{2}\sqrt{\frac{{\left(\left[A\right]+\left[T\right]+K_d\right)}^2}{4}-\left[A\right]\left[T\right]}$
where the negative root is the relevant one:

(17) $\left[\mathrm{bA}\right]=\frac{\left(\left[A\right]+\left[T\right]+K_d\right)}{2}-\sqrt{\frac{{\left(\left[A\right]+\left[T\right]+K_d\right)}^2}{4}-\left[A\right]\left[T\right]}$
Substituting concentrations for particle numbers/the number of particles per mole (C)/unit of volume (V) yields

(18) $\frac{\mathrm{bA}}{\left(\mathrm{CxV}\right)}=\frac{\left(\frac{A}{\left(\mathrm{CxV}\right)}+\frac{T}{\left(\mathrm{CxV}\right)}+K_d\right)}{2}-\sqrt{\frac{{\left(\frac{A}{\left(\mathrm{CxV}\right)}+\frac{T}{\left(\mathrm{CxV}\right)}+K_d\right)}^2}{4}-\frac{\mathrm{AT}}{{\left(\mathrm{CxV}\right)}^2}}$
and simplified to

(19) $\begin{array}{cc}\mathrm{bA}=\frac{\left(A+T+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}{2}-\sqrt{\frac{{\left(A+T+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}^2}{4}-\mathrm{AT}}& \left(\mathrm{III}\right)\end{array}$
where A=total number of anti-ligands A T=total number of ligands T V=the reaction volume (liters) C=Avogadro's constant (6,02210.sup.23 particles/mole)

(20) Given that the LMA applies to each reaction between different anti-ligands with given affinity and specificity for their respective target ligands, the number of anti-ligands bound to ligands following a selection process may be calculated by applying the LMA and equation (III).

(21) Furthermore, if there is no qualitative difference between the anti-ligands associated with the populations of subtractor or target ligands, i.e. that there is no change in the physico-chemical properties of the ligand during the method, then the number of anti-ligands that have bound to target ligands at equilibrium will be equal to the total number of bound anti-ligands multiplied by the ratio of target ligands on target ligand constructs to total ligand (subtractor and target ligand):

(22) Introducing

(23) Cp=the number of target ligand constructs C.sub.S=the number of subtractor ligand constructs T.sub.P=the number of T ligands on Cp T.sub.S=the number of T ligands on C.sub.S

(24) If target and subtractor constructs are mixed then the total number of ligands will be:
T.sub.Tot=(T.sub.PC.sub.P+T.sub.SC.sub.S)

(25) And the number of anti-ligands (A) bound to the positive constructs at equilibrium (bA.sub.P) is given by:

(26) 0$\begin{array}{cc}{\mathrm{bA}}_p=\mathrm{bAx}\frac{\left(T_p{\mathrm{xC}}_p\right)}{T_{\mathrm{Tot}}}& \left(\mathrm{IV}\right)\end{array}$

(27) Furthermore the combination of equations (III) and (IV) yields

(28) $\begin{array}{cc}{\mathrm{bA}}_p=\left\{\frac{\left(A+T+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}{2}-\sqrt{\frac{{\left(A+T+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}^2}{4}-\mathrm{AxT}}\right\}x\left\{\frac{\left(T_p{\mathrm{xC}}_p\right)}{\left(\left(T_p{\mathrm{xC}}_p\right)+\left(T_s{\mathrm{xC}}_s\right)\right)}\right\}& \left(V\right)\end{array}$

EXAMPLE 2

Optimising Ligand Concentrations

(29) The equations exemplified in example 1 show that utilisation of high concentrations of both the first subtractor ligand and the second target ligand is instrumental in the efficient retrieval of anti-ligands with specificity for low expression and differentially expressed ligands, as well as for the reduction of anti-ligands with specificity for commonly expressed ligands.

(30) Ligand concentration may be increased by several means. In all cases ligand concentration is increased by moving from two-dimensional coupling of ligand (coupling to a two-dimensional solid-phase) to use of ligand free in suspension or solution (three-dimensional).

(31) In cases where binding is dependent on the ligand being used in its native configuration, such as for cell surface ligands, then ligand concentration is maximised by increasing the ratio of ligand construct surface area to ligand construct volume.

(32) For example, cell surface antigens may be used in the form of small plasma membrane vesicles free in suspension, as opposed to using whole cells fixed to a 2-dimensional surface. This has the additional advantage of increasing the stability of the ligand in suspension or solution, thus promoting the ligand-anti-ligand equilibrium interaction.

(33) If the ligand source has a spherical (or substantially spherical) form, this is described mathematically by the following equation:
Ap/Vp=(4r.sup.2)/(4r.sup.3/3)=/3r
Where Ap=sphere area Vp=sphere volume
i.e, the smaller the radios of the sphere, the greater the ratio of ligands/volume and the more particulate (suspension like) the ligand.

EXAMPLE 3

Testing Equation III

(34) Highly diversified molecular libraries, such as phage display libraries, typically comprise some 110.sup.10 genotype unique library members. When considering a phage library, a typical 2.5 ml library stock (VolumeV) contains some 210.sup.13 total displayed proteins, meaning that each individual library member has 2000 displayed proteins, i.e. that each genotype is represented by 2000 copies. On average, some 10% of these particles (200) (Atarget specific anti-ligands) may be estimated to display ligand binding ability.

(35) Insertion of these parameters into equation III and using a desired binding affinity of a good anti-ligand (K.sub.d10 nM), the minimum number of ligands (T) needed to capture (bAbound anti-ligands) a given number of genotype specific anti-ligands may be calculated applying equation III (FIG. 1).

(36) When cells are used as the ligand source, these parameters border that of what is practically feasible using whole cells in conventional biopanning procedures. The number of cells that may be suspended in a 2 ml volume is finite, as increasing cell numbers results in an increased viscosity of the medium, shear forces and impairment of the equilibrium affinity interaction between anti-ligand and target ligand. Overcoming these problems by the methods of the invention permits isolation, of anti-ligands to low expression ligands present on only a minority of cells within a heterogeneous cell population.

(37) When attempting to isolate anti-ligands that are specific for ligand(s) that are more abundant in one ligand population compared to anothere.g. antigens on cancer versus non-cancerous cellsthere is a still greater demand on utilisation of high concentrations of antigen in a controlled manner. In a highly diversified molecular library there will, in principle, exist anti-ligands to every ligand present in the mixture. Anti-ligands specific for upregulated or unique ligands require enriching relative to common anti-ligands.

(38) Enrichment may be achieved by a competitive biopanning procedure in which ligands from the control population are added in excess.

(39) A molecular library contains anti-ligands with specificity to a large number of different ligands whose concentrations differ between two given complex ligand mixtureshere exemplified by, but not restricted to, cellsas shown in Table I.

(40) It is further assumed that unregulated or unique ligands are rare in comparison to ligands that are similarly distributed between the subtractor and target ligand cell constructs (indicated in Table I by 100 thousand-fold greater anti-ligand representatives for ligand categories TN1 and TN2).

(41) In Table I, antigens (ligand) are divided into categories based on their absolute, and relative expression between target and subtractor ligand constructs. In this example, the ligand constructs are cells, and it is assumed that two different cell populations differ in their expression of certain antigens, but also share expression of certain other antigens.

(42) Antigens belonging to categories TR1 and TR 2 are only expressed on target cell (TR=Target cell Restricted); antigen of the TE1, TE2, TE3, and TE4 categories are expressed and enriched at different level on the target cells when compared to the subtractor cells; and target antigens belonging to categories TN1.X and TN2.X are those antigens expressed at equal densities on both the target and subtractor cells, but at different absolute numbers per cell.

(43) Furthermore, it is assumed that antigens expressed at equal densities on target and subtractor cells (TN1.X and TN2.X) are 100 thousand-fold more common than either the target cell enriched (TE) or unique (TR) antigens.

(44) TABLE-US-00001 TABLE I Anti-ligand molecule categories by positive cell antigen and subtractor cell antigen prevalence Positive cell Subtractor cell Anti-ligand Anti-ligand expression expression Category Specificity (antigens/cell) T.sub.p (antigens/cell) T.sub.s R1 Antigen TR1 1,000,000 0 R2 Antigen TR2 1,000,000 0 E1 Antigen TE1 50,000 1,000 E2 Antigen TE2 50,000 10,000 E3 Antigen TE3 10,000 1,000 E4 Antigen TE4 5,000 1,000 N1.1-N1.100000 Antigen TN1.1- 100,000 100,000 TN1.100000 (100,000) (100,000) N2.1-N2.100000 Antigen TN2.1- 1,000,000 1,000,000 TN2.100000 (100,000) (100,000)

EXAMPLE 4

Testing Equation V

(45) If anti-ligands R1, R2, N1 and N2 derived from a highly diversified molecular library are mixed with their respective target antigens TR1, TR2, TN1 and TN2, and where differential expression on target and subtractor cells occurs, then the number of anti-ligands bound to positive cells at equilibrium are given by

(46) $\mathrm{bR}1p=\left\{\frac{\left(R1+\mathrm{TR}1+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}{2}-\sqrt{\frac{{\left(R1+\mathrm{TR}1+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}^2}{4}-R1\mathrm{xTR}1}\right\}x\left\{\frac{\left(\mathrm{TR}{1}_p{\mathrm{xC}}_p\right)}{\left(\left(\mathrm{TR}{1}_p{\mathrm{xC}}_p\right)+\left(\mathrm{TR}{1}_s{\mathrm{xC}}_s\right)\right)}\right\}$$\mathrm{bN}1p=\left\{\frac{\left(N1+\mathrm{TN}1+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}{2}-\sqrt{\frac{{\left(N1+\mathrm{TN}1+\left(K_d\right)x\left(\mathrm{CxV}\right)\right)}^2}{4}-N1\mathrm{xTN}1}\right\}x\left\{\frac{\left(\mathrm{TN}{1}_p{\mathrm{xC}}_p\right)}{\left(\mathrm{TN}{1}_p{\mathrm{xC}}_p\right)+\left(\mathrm{TN}{1}_s{\mathrm{xC}}_s\right))}\right\}$
and so forth.

(47) Substituting antigen numbers TR1, TR2, TE1, . . . , TE4, TN1 and TN2 with the numbers shown in Table I, and keeping all other parameters constant and as described above, the number of anti-ligands specific for different antigen categories that are bound to target cell (population) antigen at equilibrium, following competitive selection, may be calculated as a function of added target cells and subtractor cells (FIG. 2).

(48) In essence FIG. 2 demonstrates the following: Increasing addition of target ligand results in the isolation of increasing numbers of anti-ligands with specificity for: a) Target cell upregulated ligands and target cell unique ligands; and b) Lower abundant target cell unregulated ligands. Increasing addition of subtractor cells results in better subtraction (removal) of anti-ligands with specificity to ligands expressed at equal densities on target cells and subtractor cells, such that, at high enough ligand concentrations and ratios between target and subtractor cells, all anti-ligands with specificity to commonly expressed ligands will be bound to the subtractor cells (e.g., FIG. 2510.sup.7 target cells and 110.sup.10 subtractor cells).

(49) In the above example, and given the reaction parameters of FIG. 2, in order for anti-ligands with specificity for all target cell upregulated ligand (including the low expression and low unregulated antigen category E4) to be isolated, 510.sup.7 positive cells need to be utilised in the selection process.

(50) A subtractor ligand excess of two-hundred-fold (110.sup.10) will deplete all anti-ligand with specificity for commonly expressed ligands while anti-ligands with specificity for all target cell upregulated or unique ligands are retained.

(51) In a different selection process set up, where the anti-ligand copy number is not kept constant, utilisation of 5000 copies of anti-ligand (of identical or different genotype) against each ligand would require 110.sup.6 target cells and 510.sup.9 subtractor cells in order to capture anti-ligands of all desirable specificities and the removal of anti-ligands with specificity for all commonly expressed ligands.

(52) For depletion of anti-ligands with specificity for commonly expressed ligands, the ratio of target to subtractor ligand construct used in the selection process seeds to exceed the total number of bound anti-ligands with specificity for the highest commonly expressed ligands.

(53) The threshold for such depletion may be determined empirically by titrating the library member input (from a highly diversified molecular library) for the selection process. If the above criteria are met, only anti-ligands with specificity for target cell enriched or target cell unique ligands, will predominate over anti-ligands with specificity for commonly expressed ligands regardless of their affinities (FIG. 3).

(54) Low affinity anti-ligands (Kd=1 M in FIG. 3) that stick/to the cells are then able to be washed away. Due to viscosity effects, increased shearing forces, and impairment of the equilibrium reaction, between ligand and anti-ligand, the number of cells needed to fulfil these criteria exceeds that of what is feasible using whole cells, currently available molecular libraries and established biopanning techniques.

(55) As an alternative to immediate depletion of all anti-ligands with specificity for target population/subtractor population commonly expressed ligands, multiple rounds of selection may be performed to enrich for anti-ligands with specificity for ligands uniquely present or upregulated on the target cell. Depending on the quality of the anti-ligand library, the expected abundance of the targeted ligand of interest, and the availability of subtractor ligand construct, such methodology may prove useful.

(56) The ratios between anti-ligands with different specificity increase exponentially with the number of selection rounds. If the total concentration of target ligand (A) is much greater than the concentration of free anti-ligand molecules (fT) ([A]>>>[fT]), as is the case in any given selection process utilising highly diverse molecular libraries, then the concentration of free ligand can be replaced by the total concentration of ligand ([fT][T]) yielding kD[fA][T]/[bA]. Replacing [fA] by [A][bA] yields Kd([A][bA])[T]/[bA] and rearranged to [bA][A][T]/(Kd[T]).

(57) Replacing concentrations for particle numbers/number of particles per mole(C)/volume (V), the number of captured anti-ligands is obtained as bAAT/(CVKd+T).

(58) If target cells are mixed with subtractor cells, the number of anti-ligands captured on target cells (bA.sub.P) is given by bA.sub.PbA(T.sub.PC.sub.P)/T.sub.TotAT.sub.Tot/(CVKd+T.sub.Tot)(T.sub.PC.sub.P)/T.sub.TotA/(CVKd+T.sub.Tol)(T.sub.PC.sub.P).

(59) $\begin{array}{cc}{\mathrm{bA}}_p\left(\frac{{\mathrm{AxT}}_{\mathrm{Tot}}}{\mathrm{CxVx}\left(K_d+T_{\mathrm{Tot}}\right)}\right)x\left(\frac{T_p{\mathrm{xC}}_p}{\left(T_p{\mathrm{xC}}^p\right)+\left(T_s{\mathrm{xC}}_s\right)}\right)& \mathrm{VI}\end{array}$

(60) If:

(61) $\left(\frac{T_{\mathrm{Tot}}}{\mathrm{CxVx}\left(K_d+T_{\mathrm{Tot}}\right)}\right)x\left(\frac{T_p{\mathrm{xC}}_p}{\left(T_p{\mathrm{xC}}^p\right)+\left(T_s{\mathrm{xC}}_s\right)}\right)$
is constant between selection rounds (reaction parameters other than anti-ligand input (A.sub.in)) and equal to for anti-ligand.sub.Aligand.sub.A interaction and for anti-ligand.sub.Bligand.sub.B interaction, then the number of captured anti-ligands in selection 1 specific for ligand.sub.A is given by a1=A.sub.in, and captured anti-ligands specific for ligand.sub.B is given by b1=B.sub.in.

(62) Similarly, following x-fold amplification of eluted anti-ligand the number of anti-ligands captured on target cell ligands A, or B In selection round 2 is given by a2=x*a1=x*A.sub.in and b2=x*b1=x*B.sub.in.

(63) In selection round 1 the ratio of captured anti-ligands specific for ligand.sub.A to captured anti-ligands specific for Uganda is givers by (A.sub.in)/(B.sub.in). In selection two the ratio will be (A.sub.in)/(B.sub.in), and after n selections the ratio will be (A.sub.in/B.sub.in)(/).sup.n.

(64) Hence, with increasing rounds of panning, anti-ligands- with specificity tor the target cell unique or upregulated ligand(s) will prevail over anti-ligands with specificity to commonly expressed ligands.

(65) However, multiple panning rounds also decrease the frequency of anti-ligands on the target ligand constructs with specificity for low expression target cell unique or enriched ligands (in comparison to anti-ligands with specificity for highly expressed target cell unique or enriched ligands) e.g. ratios of anti-ligands from categories E4 to R1 increase from 1:44 to 1:25000 after 3 rounds of panning when 10.sup.8 target cells and 10.sup.9 subtractor cells are utilised (FIG. 4). This makes the screening process for such anti-ligands extensive, and thereby further demonstrating the need for utilisation of ligand concentrations as high as possible in the selection process.

(66) Multiple selection rounds are performed using reaction parameters other than anti-ligand input (volume, number of target cells and subtractor cells) that are kept constant (FIG. 4).

(67) If large enough concentrations of ligand are utilised all anti-ligands possessing, e.g., nanomolar levels of binding affinity will be bound to their respective ligand. With reaction parameters as used in the previous examples it would take some 710.sup.15 ligand to achieve this.

(68) If the lowest abundant ligand population of interest is present at such levels, and if the subtractor ligand population is added at an excess compared to the target ligand population exceeding the number of anti-ligands completed at equilibrium (200 in this case), anti-ligands specific for ligands that are more abundant in the original target cell population compared to the subtractor cell population will be isolated in numbers in line with the increase in target cell population compared to the subtractor population in a target ligand population abundance independent manner.

(69) Anti-ligands with specificity for ligands found at identical numbers in original target and subtractor ligand populations will be bound to subtractor ligand (Table II/FIG. 5).

(70) TABLE-US-00002 TABLE II Anti-ligand molecule categories by positive antigen population and negative antigen population prevalence Positive antigen Negative antigen population population abundance abundance Anti-ligand Anti-ligand (molecules (molecules Category Specificity 3.5 10.sup.14)T.sub.p 3.5 10.sup.14)T.sub.s Positive population 1.00E+06 0.00E+00 restricted antigen A Positive population 1.00E+06 1.00E+04 enriched antigen B Positive population 1.00E+05 1.00E+03 enriched antigen C Positive population 1.00E+06 1.00E+05 enriched antigen D Positive population 1.00E+05 1.00E+04 enriched antigen E Positive population 5.00E+00 1.00E+00 enriched antigen Positive and negative 1.00E+06 1.00E+06 population commonly abundant antigens (highly abundant) Positive and negative 1.00E+02 1.00E+02 population commonly abundant antigens (rare)

EXAMPLE 5

Preferred Embodiment

(71) In a preferred aspect the invention is used to isolate anti-ligands with specificity for cell surface antigens in their native configuration and independent of their nature (protein, carbohydrate, lipid, complex). Additionally the antigens being bound are those upregulated or uniquely expressed on one cell type compared to another (e.g. transformed cancer cell, viral/microbial/parasite/fungal infected cell or other agonist stimulated or infection activated cell versus control cells), (FIGS. 2, 4 and 5)

(72) When utilised for selection of antibody derived anti-ligands (e.g. scFv-, Fab-, or Fv-encoding anti-ligands), the method, simultaneously with the screening process, generates therapeutic antibody candidates that react with target antigen in its native configuration at the cell membrane.

(73) Because such large concentrations of antigen are needed, antigen is used in a form that does not impair the equilibrium reaction. Therefore, antigen is used in forms that occupy minimal space and impose little increase in viscosity and shearing forces.

(74) For example, when anti-ligands to cell surface antigens are sought, a competition biopanning process utilising target whole cells and excess subtractor cell membranes mixed with members of a highly diversified, molecular anti-ligand library may be used, followed by density separation on a Ficoll or Percoll/bovine serum albumin gradient and selective isolation of target cells and anti-ligands specific for target cell upregulated and unique antigens.

(75) In this methodology the target ligand (antigen) population is in the form of whole cells (high density) and the subtractor ligand (antigen) is in the form of plasma membrane vesicles or enucleated cells (low density).

(76) The target and subtractor antigen populations are mixed with members of a highly diverse molecular library in a controlled manner based on the equations described herein.

(77) For example, 510.sup.7 target whole cells are mixed with cell membrane vesicles of 110.sup.10 subtractor cells and mixed with members from a highly diversified library at an anti-ligand specific copy number of 200 (typically producing anti-ligands of Kd=10.sup.8 M when selecting on pure antigen), one can expect to isolate anti-ligands specific for 10-fold or greater upregulated antigens including those exposed at such low densities as 10,000 per target cell.

(78) The reaction is incubated to reach equilibrium. Following competitive biopanning, library members bound to the target population are separated from unbound anti-ligands and those anti-ligands bound to control subtractor antigen by density centrifugation separation, resulting in enrichment of phage specific for highly expressed antigens present among the studied population.

(79) Where the desired target antigen expression is higher in the subtractor population the process is reversed, so that the subtractor ligand population becomes target ligand population and vice versa.

(80) Besides generating anti-ligands with specificity for differentially expressed and unique ligands, use of different density separation means on a density gradient, offers several advantages including: Physical and spatial separation of anti-ligands complexed to positive ligand from unbound anti-ligands and anti-ligands with specificity for ligand found in the control population. Ficoll washing increases shear force. Hence, such washing is more efficient and less washing repetitions (panning rounds) are needed; and there is minimal dissociation of specifically bound (higher affinity) anti-ligands of interest. Does not require tagging or chemical modification of cells (compare FACS (fluorescence activated cell sorter) or MACS (magnetic activated cell sorter) based competitive biopanning) that might alter cell surface ligand configuration/conformation and/or composition.

EXAMPLE 6

All Membrane Vesicles as Separation Means

(81) Whole cells can be replaced by membrane vesicles produced in a higher density media, allowing for even higher concentrations of ligand to be utilized without compromising the equilibrium reaction.

EXAMPLE 7

Testing the Effects of Stimuli on Ligand Up-/Down-regulation

(82) A further embodiment of the invention may be used to isolate anti-ligands with specificity for cellular ligands that are expressed at a very low density in only a small number of cells within the cell population being studied.

(83) For example, a certain stimulus may be suspected to trigger the upregulation or downregulation of a cell surface antigen present on an unknown cell subpopulation present in blood.

(84) Cells derived from whole blood exposed to this stimulus may be mixed with plasma membranes derived from whole blood prior to exposure to the stimulus, and a competitive biopanning reaction analogous to that described above.

EXAMPLE 8

Diagnostic Use of the Screening Method

(85) A further example of the invention, allows anti-ligands against ligands present at different abundance in biological samples (e.g. plasma, urine, cerebrospinal fluid) to be isolated from highly diversified molecular libraries. Such anti-ligands may subsequently be used for, e.g., protein expression analysis and identification of potential biomarkers.

(86) If sufficiently high concentrations of ligands are used, the method of the invention allows for selective isolation of anti-ligands against up-regulated or unique ligands when comparing protein composition in two different samples (FIG. 2). This ultimately allows for isolation of anti-ligands specific for ligands that are more abundant in one population compared to another population, in a manner that is independent of the relative ligand concentrations within the positive ligand population (FIG. 5).

(87) Due to the extreme concentrations of ligand needed to accomplish the latter, ligand should preferably be used in suspension or solution. For example, target population ligand can be split and tagged at several different positions to minimise destruction and eradication of relevant ligands, while subtractor population ligands can be used, untagged, or mock treated. Tagging of the positive ligand population provides a means for subsequent retrieval of positive population ligands and binders bound to positive population ligands only by use of tagged ligand completed with e.g. counter tagged magnetic beads.

(88) An application of this method would be to pool plasma samples from a population of patients with a certain illness and compare to a plasma samples from a control population. In this case the patient plasma samples would be split and tagged, and the control population would be untagged (see example 11).

EXAMPLE 9

Experimental Testing of Equation(s)

(89) This example demonstrates the experimental use of the preferred embodimentdensity competition biopanning using whole cells and cell membrane vesicles to isolate anti-ligands with specificity for a target cell unique ligand.

(90) The model phage pKBitCD40L-2 is a CD40L specific phage and the model phage pBitBLTR-2 is a general mouse fibroblast anti-ligand phage [BioInvent, Sweden].

(91) In order to demonstrate efficient subtraction of phage with specificity for common antigens we used the pBitBLTR-2 phage. This phage was isolated in a previous selection process for its general ability to bind non-transfected, CD32 transfected and CD40L transfected mouse 3T3 fibroblast cells.

(92) FIG. 6 shows the number of kanamycin resistance carrying phage (CD40L specific) and ampicillin resistance carrying phage (mouse fibroblast binder) that were elated from CD40L transfected 3T3 cells (510.sup.6). CD32transfected 3T3 cells (2.610.sup.6), or non-transfected 3T3 cells (610.sup.6) following centrifugation based biopanning using a 900 l phage stock containing 10.sup.6 pfu CD40L specific phage, 10.sup.6 pfu mouse fibroblast general binder phage and 10.sup.10 pfu R408 helper phage (no antibiotic resistance).

(93) The affinity of the CD40L specific phage scFv was previously demonstrated to be K.sub.d=810.sup.9M. Mouse fibroblast expression of CD40L and CD32 was confirmed by staining with monoclonal antibodies (MAb) and flow-cytometric analysis (not shown).

(94) CD40L anti-ligand specificity of the kanamycin resistant phage was confirmed by selective retrieval of these phage from CD40L-transfected mouse fibroblasts. The general mouse fibroblast specificity of the ampiciliin resistance carrying phage was confirmed by the retrieval of similar numbers of ampicillin resistance carrying phage following elution from CD40L-transfected, CD32-transfected, and non-transfected mouse fibroblasts (FIGS. 7A & 7B).

(95) The efficacy of utilising excess membrane as subtractors of phage with irrelevant specificity was demonstrated by comparing the ratios of kanamycin and ampiciliin resistant phage elated from 1) a selection process with CD40L-transfected cells only, 2) a pre-selection process utilising whole CD32 transfected mouse fibroblasts at an excess of fourfold; and 3) competition biopanning using a fifty-fold excess of CD32-transfected plasma membranes.

(96) Plasma membrane competition was five-fold more efficient than preselection using whole cells, and utilisation of membranes resulted in a less viscous selection reaction compared to when whole cells were used.

(97) Experimental Methods

(98) Phage Manufacture

(99) An antibody library of human scFv fragments was used to isolate scFv antibody fragments recognising the CD40 ligand antigen (CD40L) on CD40L-transfected mouse 3T3 cells.

(100) In the first pre-selection round the library (710.sup.13 CFU/ml, 1.8 ml) was incubated with 9.310.sup.6 un-transfected mouse 3T3 fibroblast cells (1 ml), and 0.2 ml R10 medium {RPMI 1640, 10% FCS, 1 non-essential amino acids, 2 mM BDTA, 50 mM HEPES, (GIBCO BRL, Gaithersburg, Md. 20877, USA)} for 75 min at 15 C. with slow rotation and a total reaction volume of 4.0 ml.

(101) The cells were centrifuged for 10 min at 1500 rpm, 15 C., and the phage containing supernatant transferred to a second pre-selection step analogous to the first, except that CD32-transfected mouse fibroblasts (110.sup.7) were used.

(102) Following pre-selection, phage were precipitated by treatment with polyethylene glycol (PEG), resuspended in 1 ml R10 medium, and were used for positive selection on 110.sup.7 CD40L-transfected 3T3 cells. Positive selection was undertaken at 20 C., during slow rotation, for 3 h.

(103) Cells were washed twice in density gradients consisting of 40% Ficoll/ 2% BSA/Dulbecco's PBS (Gibco BRL), and once in Dulbecco's PBS only.

(104) Bound phage were eluted from cells by re-suspension in 1 ml Glycine/HCl pH 2.2, at room temperature for 15 minutes, followed by neutralisation in 2M Tris-HCl (pH 7.4).

(105) Following centrifugation and collection of eluted phage, the remaining bound phage were eluted in 1 ml 0.4EDTA-trypsin solution (Gibco BRL) in R10 medium. Pre-selection and positive selection was repeated three times as above.

(106) The specificity of the selected phage scFv was determined by, firstly, screening individual phage clones for reactivity against CD32 transfected or CD40L transfected mouse 3T3 fibroblasts in cell ELISA. Some 20% of these clones showed selective reactivity with CD40L-transfected cells and were picked for further analysis.

(107) The selected clones were converted to scFv format by excision of the gene III fragment and scFv molecules were produced in the nonsuppressor strain E. coli HB101. The specificity of the clones was confirmed by scFv whole cell ELISA, and flow-cytometry, using untransfected, CD32-transfected, and CD40L-transfected mouse 3T3 fibroblast cells.

(108) For flow-cytometric analysts scFv binding was detected by biotinylated anti-flag M2 antibody (Sigma Chemical Co, St Louis, Mo., USA) and subsequent Streptavidin-PE (DAKO, Glostrup, Denmark) addition. The affinity of clone 7E for CD40L antigen was Kd=8109M as determined by scatchard blot using whole CD40L-transfected 3T3 cells as an antigenic source (unpublished data).

(109) Selection of the general mouse pBitBLTR-2 phage was carried out essentially as for pKBitCD40L-2, except BLTR-transfected 3T3 cells were used for positive selection. Clone BLTR-2 was selected as a clone that showed equal binding to non-transfected, CD32-transfected, and CD40L-transfected 3T3 cells, as determined by flow-cytometry at the scFv level.

(110) Competitive Bio-Panning

(111) Phage Stock

(112) 10.sup.6 cfu of CD40L specific phage (Kan.sup.1) 10.sup.6 cfu of mouse fibroblast general binder phage (Amp.sup.1) 10.sup.10 cfu of R408 helper phage (no antibiotic resistance)
CellsPositive (Second Target Ligand Construct) A1. CD40L transfected mouse 3T3 fibroblasts 510.sup.6 cells C1. 510.sup.6 CD40L transfected mouse 3T3 fibroblasts only C2. 2.510.sup.6 CD32 transfected mouse 3T3 fibroblasts only (limited) C3. 510.sup.6 cells non-transfected mouse 3T3 fibroblasts only
CellsNegative (First Subtractor Ligand Construct) A1. CD32 transfected mouse 3T3 fibroblast membranes equivalent to 2.510.sup.8 cells.
Method 1. A 1 ml phage stock was pre-warmed at 3 C. for 15 min (A 4 ml phagestock was prepared by diluting CD40L specific phage and BLTR general mouse 3T3 cell anti-ligand phage to 10.sup.6 cfu/ml and R408 helper phage to 10.sup.10 cfu/ml in 2% milk PBS. This phage stock was split to 41 ml, transferred to 1.5 ml Eppendorf tubes and was used to perform competitive selections A and B, and conventional pre-selection/selections A and B). The phage stock was vortexed intermittently and centrifuged for 15 min at full speed in an Eppendorf centrifuge. Where a precipitate had formed, the supernatant was transferred to a new Eppendorf tube. Skimmed milk was added to a concentration of 2%. 2. Naive CD32transfected 3T3 mouse fibroblast cell plasma membrane preparations were thawed, from 510.sup.8 cells on ice. 10 l was saved for protein concentration determination. Phage stock was re-suspended and mixed with a pipette. The stock was then incubated for 5 min on ice. 3. 510.sup.6 CD40L-transfected 3T3 or non-transfected 3T3 cells were centrifuged at 1300 rpm for 6 min at 4 C. 4. The supernatant was discarded and the CD40-transfected 3T3 or non-transfected 3T3 cells were re-suspended in the milk-phage-negative cell membrane stock solution from (step 2) and incubated at 10 C. whilst undergoing slow (end-over-end) for 4 hours. 5. The cell/cell membrane/phage incubate was transferred to a 15 ml Falcon tube containing 10 ml 40% Ficoll-Paque Plus and 2% BSA in PBS (with Ca and Mg) and centrifuged at 1500 rpm for 10 min at 4 C. 6. The supernatant was carefully removed by aspirating the uppermost supernatant first (saving the membrane fraction separate (100 l) for subsequent titration), and then a sequential working down towards the cell pellet containing bound target phage particles. As much supernatant was removed as possible (saving 25 l supernatant for titration) and the cell pellet re-suspended in 500 l of PBS-2% BSA. The wash performed in (step 5) is repeated once (saving the supernatant for titration). 7. The cells are re-suspended in 1 ml PBS, transferred to a new 15 ml Eppendorf tube and centrifuged at 1260 rpm for 10 min at 4 C. The supernatant is removed rising a pipette (saving the supernatant for titration). 8. Phage were elated from cells by the addition of 150 l of 76 mM citric acid (pH2.5) in PBS followed by incubation at room temperature for 5 min. The mixture is neutralised by addition of 200 l of 1M Tris-HCl, pH 7.4. Centrifugation is repeated and the eluted phage saved. 9. The cells are re-suspended in 1 ml trypsin, transferred to a new tube and incubated for 1.0 min. They are then inactivate with 40 l 1 mg/ml aprotinin and centrifuged, saving the supernatant for titration. The bacterial pellet is saved for infection of bacteria.
Preselection Biopanning
Phage Stock 10.sup.6 cfu of CD40L specific phage (Kan.sup.1) 10.sup.6 cfu of mouse fibroblast general binder phage (Amp.sup.1) 10.sup.10 cfu of R408 helper phage (no antibiotic resistance)

(113) Cells Positive (Target)CD40L Transfected Mouse 3T3 Fibroblasts 510.sup.6 Cells

(114) Cells Negative (Subtractor)CD32 Transfected Mouse 3T3 Fibroblasts 210.sup.7 Cells

(115) Method

(116) 1. A 1 ml phage stock was pre-warmed at 37 C. for 15 min and vortexed intermittently. The phage stock was centrifuged for 15 min at full speed in eppendorf centrifuge. Where a precipitate has formed, supernatant was transferred to new eppendorf tube and skimmed milk added to a concentration of 2%. 2. 210.sup.7 non-transfected 3T3 cells were centrifuged at 1300 rpm, 6 min, 4 C. The supernatant is discarded and the cells re-suspended by the addition of phage stock and by mixing with a 1 ml pipette. This is then incubated at 10 C. undergoing slow (end-over-end) rotation for 4 hours. Then the cells are centrifuged at 1300 rpm for 6 min at 4 C., and the supernatant transferred to a new tube. 20 l of supernatant is saved for phage titration. 3. 510.sup.6 CD40L-transfected 3T3 cells were detached, washed and centrifuged at 1300 rpm for 6 min at 4 C. 4. The supernatant was discarded, and the OD40L-transfected 3T3 cells re-suspended in the milk-phage stock solution from (step 2). This was then incubated at 10 C. undergoing slow (end-over-end) rotation for 4 hours. 5. The cell/phage incubate is transferred to a 15 ml Falcon tube containing 10 ml 40% Ficoll-Paque Plus and 2% BSA in PBS (with Ca and Mg) and then centrifuged at 1500 rpm for 10 mm at 4 C. 6. The supernatant was carefully removed by aspiration of the uppermost supernatant, first then sequential removal working down to the cell pellet containing bound target phage particles. As much supernatant as possible was removed (and saved) and the cell pellet re-suspended in 500 l of PBS-2% BSA. The wash as performed in (step 5) was repeat once (and saved for titration). 7. Cells were re-suspended in 1 ml PBS and transferred to a new 15 ml Eppendorf tube. They were then centrifuged at 1260 rpm for 10 min at 4 C. The supernatant was removed using a pipette (and saved for titration). 8. Phage were eluted from cells by addition of 150 l of 76 mM citric acid (pH2.5) in PBS followed by incubation at room temperature for 5 min. The mixture was neutralised by the addition of 200 l of 1M Tris-HCl, pH 7.4 and then centrifuged and the eluted phage saved. 9. The cells were re-suspended in 1 ml trypsin, transferred to a new tube and incubated for 10 min. They were then inactivated with 40 l 1 mg/ml aprotinin. The cells are centrifuged and the supernatant and pellet saved for infection of bacteria.
Procedure for Phage Simplification from Bacteria (10 ml) 1. Each expression is injected with 10 ml LB (100 g/ml ampicillin, 15 g/ml tetracycline with 5 l from the glycerol stocks) 2. They were grown at 37 C. and 175 rpm until OD.sub.600=0.5 and 610.sup.9 PFU of R408 helper phage per ml culture added and incubated for 30 minutes at 37 C. and 50 rpm. 3. IPTG solution was added to a final concentration of 100 M and incubated overnight at 25 C. and 175 rpm. 4. The following day the phage were harvested (as below).
Harvest PEG Precipitation of Amplified Phage Stocks 1. Bacteria were pelleted by centrifugation for 10 min at 2100 g/(3000 rpm, in Beckman GS-6) at room temperature. 2. The supernatant was sterile filtered through 0.2 m sterile filter. Any supernatants stemming from the same selection were pooled. 3. Phage were precipitated by adding volume of phage precipitation buffer (20% PEG 6000, 2.5 M NaCl) and were then incubated for at least 4 hours at 4 C. (Although this can be left incubating for weeks). 4. The preparation is centrifuged for 30 min at 4 C. and 13000 g. 5. The pellet is completely re-suspended in 500 l PBS, (incubation at 37 C. with agitation for 1 hour) and then stored at 4 C. over night.
Titration of Phages Pools 1. The phage stock was diluted in sterile PBS. (Suitable dilutions for eluted stocks 10, 10.sup.2, 10.sup.3, 10.sup.4, and for start stocks or amplified stocks 10.sup.610.sup.7, 10.sup.810.sup.9). 2. 100 l indicator bacteria (grown to OD.sub.600=0.5) were mixed with 10 of each phage dilution and incubated for 30 minutes at 37 C. and 50 rpm. 3. They were then plated on LB agar plates (100 g/ml ampicillin, 15 g/ml tetracycline, 1% glucose) and incubated overnight at 37 C. 4. The colonies were then counted and the litre calculated (CPU/ml), (CPUdilution100).
Plasma membranes Preparation from CD32-Transfected Mouse 3T3Cells 1. Cells were grown to confluency in large (500 cm.sup.2) plates. Cells (110.sup.9) were washed once with versene (50 ml PBS without calcium and magnesium containing 2 mM EDTA), and were then incubated with a small volume of versene (barely covering the cell mat) at 37 C. in a cell incubator. 2. Cells were loosened by hitting the flask against the palm of a hand and then re-suspended in PBS (with Ca and Mg), centrifuged in 50 ml tubes for 7 min at 1200 rpm 4 C., and washed again in PBS (with Ca and Mg). 3. Cells were washed once, re-suspended in approximately 14 ml 0.145 M NaCl and transferred to a 15 ml test tube. They were then centrifuged for 7 min at 1200 rpm, aspirated and re-suspended in 14 ml 0.145 M NaCl. Further centrifugation for 7 min at 1200 rpm was performed and the supernatant removed and the cell pellets frozen at 80 C. 4. The pellet was re-suspended by vortexing in 8 ml (non-transfected) or 4 ml (CD32 cells) of ice-cold Buffer A (10 mM Tris, 1 mM EDTA, 0.25 M sucrose, pH 7.0) freshly mixed with PMSF (phenylmethylsulphonyl fluoride) to a final concentration of 1 mM. 5. The cell suspension was homogenised by 20 strokes of a Dounce homogeniser (type B, pre-cooled) and the homogenate centrifuged in an Eppendorf centrifuge for 10 min at 6000 rpm at 4 C. 6. The supernatant was collected with a Pasteur pipette and the pellet re-suspended in 4 ml (non-transfected) or 2 ml (CD32 cells) of Buffer A & PMSF, then homogenised and centrifuged as above. The supernatants were combined and the pellet discarded. 7. The supernatants were layered using a Pasteur pipette, on top of 0.5 ml Cushion Buffer (Buffer C containing 37% of final solution sucrose) and centrifuged for 60 min at 4 C. using rotor SW 60.1 (pre-cooled) at 30,000 rpm (100/000g). They were balanced carefully for this centrifugation step (to <0.0005 g difference between paired samples). 8. The Membrane layer (opalescent) was collected at the cushion interface, where it was tried to get all the material in a small amount of liquid. The collected material was then diluted with 4 vol. of cold Buffer C (10 mM Tris, 1 mM EDTA, pH 7.0 ) and centrifuged at 30,000 g (rotor JA 25-50, 20,000 rpm) for 30 min at 4 C. The resulting pellet was re-suspended by vortexing in 150 l of cold Buffer C. The resuspended mixture was frozen and stored at 80 C. 9. Membrane protein concentration was determined by BCA kit (Pierce Biotechnology, Rockford, Ill., USA) to 60 mg/ml.

EXAMPLE 10

Further Experimental Testing of the Equation(s)

(117) This example demonstrates a further experimental use of how the present invention may be used to isolate, from a highly diversified scFv phage library, binders that are specific for cell surface antigens uniquely distributed on one cell population.

(118) The binder hit-rate following three rounds of competitive selection utilising the present invention is shown in comparison to that of a conventional negative pre-selection with a subsequent positive-negative selection (i.e. three rounds of selection).

(119) B lymphocytes express a number of cell surface antigens that are not normally found on T lymphocytes. Such antigens include the B cell immunoglobulin receptor and co-receptors thereof (CD19, CD21), the 4 transmembrane spanning ion channel CD20, and the Fc receptor CD32.

(120) B cells, however, express cell surface antigens that are also found common to T lymphocytes, including integrin receptors e.g. LFA-1 (CD11a/CD18), VLA-2, and VLA-4, ICAM-1, complement deactivating receptors e.g. DAF and Protectin, and cytokine receptors like IFN-R and TNFR. Hence, B and T cells provide an attractive model system to test the applicability of competitive biopanning from a phage scFv display library.

(121) In this experiment 210.sup.13 phage particles from the highly diversified n-CoDeR library comprising some 10.sup.10 genotype unique binders are mixed with whole B lymphoma cell line Ramos cells (positive selection), and plasma membrane or crude membrane vesicles from the T leukaemia cell line Jurkat (negative selection). Binders specific for antigens that are uniquely expressed on the B lymphoma cell line Ramos, compared to the T cell leukaemia cell line Jurkat, are to be selectively isolated.

(122) Positive and Negative Cell Number Calculation for Selection

(123) Cell numbers to be used in the different selections round were calculated using equation VI. Reaction parameters used for calculations were as shown in Table III (IIIA & IIIB).

(124) TABLE-US-00003 TABLE III IIIA Competition selection Antigen Expression (Antigens/cell) Positive cell Subtractor cell Phagemid specificity expression expression + Cell (B Cell) Resticted Antigen 1.00E+05 0.00E+00 + Cell (B Cell) Enhanced Antigen 1.00E+05 1.00E+03 + Cell (B Cell) Enhanced Antigen 1.00E+05 1.00E+04 + Cell (B Cell) Enhanced Antigen 5.00E+04 1.00E+03 + Cell (B Cell) Enhanced Antigen 5.00E+04 1.00E+04 + Cell (B Cell) Enhanced Antigen 1.00E+04 1.00E+03 (B Cell/T Cell) Commonly expressed 1.00E+06 1.00E+06 antigen (highly expressed) (B Cell/T Cell) Commonly expressed 1.00E+05 1.00E+05 antigen (moderately expressed) cell (T cell) restricted Antigen 0 1.00E+05 Non-binder 0 0 AF(R1-R2) 1.50E+03 AF(R2-R3) 6.00E+04 Kd 1.00E08 Avogadro's Constant 6.02E+23 Genotype specific phage input R1 2.00E+02 Reaction volume V(L) R1 2.00E03 Reaction volume V(L) R2 5.00E04 Reaction volume V(L) R3 5.00E04 Number of Cell + cells (B cells) equivalents used in (T cell mem- selection rounds brane vesicles) R1 5.00E+07 2.00E+09 R2 5.00E+06 1.00E+09 R3 5.00E+06 1.00E+09 IIIB Phage particles recovered on positive cell Phage Relative input Recovered* frequency R1 2.00E+02 59 0.296 2.00E+02 53 0.264 2.00E+02 27 0.133 2.00E+02 30 0.152 2.00E+02 14 0.073 2.00E+02 7 0.035 2.00E+02 5 0.024 2.00E+02 5 0.023 2.00E+02 0 0.000 2.00E+02 0 0.000 R2 9.E+04 1.3E+04 0.482 8.E+04 8.7E+03 0.336 4.E+04 1.5E+03 0.953 5.E+04 2.7E+03 0.102 2.E+04 4.1E+02 0.016 1.E+04 1.3E+02 0.005 7.E+03 3.6E+01 0.001 7.E+03 3.3E+01 0.001 0.E+00 0.0E+01 0.000 0.E+00 0.0E+00 0.000 R3 8.E+08 1.1E+08 0.60022 5.E+08 5.8E+07 0.32542 9.E+07 3.3E+06 0.01855 2.E+08 9.4E+06 0.05248 2.E+07 4.6E+05 0.00259 8.E+06 9.4E+04 0.00053 2.E+06 1.1E+04 0.00006 2.E+06 9.7E+03 0.00005 0.E+00 0.0E+00 0.00000 0.E+00 0.0E+00 0.00000 Table III Selection reaction parameters used in the whole Cell (Ramos)/membrane vesicle (Jurkat) competitive biopanning. Kd = lowest relevant affinity of target ligand specific anti-ligands to be isolated, genotype specific phage input R1 = anti-ligand (phage scFv copy number of starting anti-ligand library), AF = amplification factor used to calculate phage input into selection rounds R2 and R3 following elution and amplification of phage rounds R1 and R2 respectively. Relative frequency = the expected frequency of isolated anti-ligands with specificity for antigen belonging to one antigen category compared to all anti-ligands isolated.

(125) Positive and negative cell numbers were chosen such that, following three rounds of selection, binders with specificity for antigens expressed uniquely on B cells will be 10,000-fold over an antigen expressed at equal density on B and T cells.

(126) The input number of phage binders specific for different categories of antigen (positive cell enriched, positive cell unique, or positive/negative cell commonly expressed antigen) in selection rounds 2 and 3 was calculated by multiplying the calculated number of eluted phage, specific for different categories of antigen following selection rounds 1 and 2, with the amplification factor (AF).

(127) The amplification factor was obtained by dividing total number of amplified phage following the relevant selection round with the total number of eluted phage from the same selection round.

(128) FIGS. 8A and 8B show the comparison of biopanning by the present method and by previously known methods on B cell (Ramos) versus T cell (Jurkat) scFv clones. The B cell reactivity and specificity of ScFv's isolated using the invention is dramatically enhance compared to that of ScFv selected by conventional biopanning.

(129) Experimental Methods

(130) Cell Captures

(131) The Jurkat T cell line, clone E6-1, and the Ramos B lymphoma cell line, were obtained from ATCC and cultured in RPMI 1640 supplemented with 10% FCS (heat-inactivated for Ramos Cells only), 10 mM HEPES and 1 mM Sodium pyruvate, in a humidified atmosphere at 37 C. The cells were maintained at 1-210.sup.6 cells/ml (<110.sup.6 cells/ml for Jurkat).

(132) Jurkat T cell Plasma Membrane Preparation

(133) Jurkat Cell Culture

(134) Jurkat E6-1 cells were maintained in RPMI-1640 with Glutamax I (Gibco, #61870-010) supplemented with 10% foetal calf serum (Gibco, Lot no 1128016) 1 mM Sodium pyruvate (Gibco) and 10 mM Hepes buffer (Gibco) in a humidified atmosphere of 5% CO.sub.2 at 37 C., and at cellular densities between 110.sup.5 to 110.sup.6 cells/ml. In the final passage, cells were allowed to reach a maximal density of 210.sup.6 at which point they were harvested.

(135) Cell Disruption

(136) 1. Cells were harvested from culture by centrifugation in 500 ml Centrifuge tubes (Corning, #431123) placed in tube adapters, 1500 rpm, 15 min at 4 C. 2. The supernatant was discarded and washed in 0.145M NaCl. Cell suspensions were pooled, cells counted (510.sup.9 cells total), and centrifugation repeated. 3. Cell disruption was performed by hypo-osmotic shock in 1 mM NaHCO.sub.3 1.5 mM MgAc pH 7.4 on ice for 10-30 min and subsequent nitrogen cavitation occurred in a Yeda press, 40 bar (4000 kPa) for 15 min at 0 C. Cell concentration did not exceed 510.sup.7 cells/ml. 4. Following disruption 150 l 0.5 M EDTA was added to the homogenate suspension to yield a final EDTA concentration of 1 mM (addition of EDTA prevents aggregation of membrane vesicles). 5. A) Crude membrane isolation: The homogenate (50 ml) was centrifuged for 10 min at 1900 g (4000 rpm in a SS34 rotor) to remove unbroken cells and nuclei, and the supernatant collected. Washing and re-centrifugation of pellet was avoided, as the fragile nuclei tend to disrupt, causing DNA leakage and aggregation; or B) Plasma membrane isolation: 10 ml of 37.2% sucrose was layered at the bottom of 638.5 ml Beckman ultra centrifugation tube, and 627 ml of the cell homogenate from step 2 above was carefully layered on top. The tube was centrifuged at 27000 rpm in a swing-out SW28 rotor (639 ml nominal capacity) for 2 h 45 min at 4 C. Plasma membranes were isolated from the tubes as the white band of the interphase between the sucrose cushion and the sample phase, and PM were pooled, split between 435 ml tubes and dilated in TE buffer (1 mM Tris/0.25 M sucrose/0.25M EDTA buffer) to a total volume of 35 ml. 6. Ultra-centrifugation was in a Beckman Type 45.Ti rotor (nominal capacity 694 ml Nalgene tubes) at 40.000 rpm (approx. 200.000g) for 1 h at 4 C. 7. The supernatants were discarded and any remaining buffer was removed using a 1 ml Finnpipette. The plasma membrane pellets were scraped off the bottom of tubes with a metal bar, and transferred to a small dounce homogeniser. Pelleted membranes were re-suspended by homogenisation in a total volume of 2.5 ml TE-buffer containing 10 mM Hepes (10 mM Hepes/1 mM Tris/0.25M sucrose/0.25M EDTA buffer) by 5-10 strokes with a loose fitting Dounce glass piston. Approximately, membranes derived from some 210.sup.9 Jurkat cells can be resuspended per ml of resuspension (TE) buffer.
Protein Concentration Determination

(137) Protein concentration determination was performed using the BCA kit according to the manufacturer's instructions. Briefly, a double BSA standard was prepared by 2-fold dilutions (10 l sample+10 l buffer) in PBS of a 2 mg/ml BSA stock solution. A standard curve was generated and used to determine the total protein concentration of membrane samples.

(138) Plasma Membrane Activity (By Alkaline Phosphatase Assay)

(139) Alkaline Phosphatase Solutions

(140) Substrate Solution

(141) 1 tablet p-NPP per 10 ml borate buffer (1.5 mg/ml final concentration) in 50 mM sodium borate buffer (pH 9.8), 1.0 mM MgCl.sub.2

(142) Triplicate samples were diluted in Borate/MgCl2 buffer by transferring 50 l sample to 50 l dilution buffer (50 mM sodium borate buffer (pH 9.8), 1.0 mM MgCl.sub.2). 200 l substrate solution (1 tablet p-NPP per 10ml borate buffer to 1.5 mg/ml final concentration in 50 mM sodium borate buffer, pH 9.8, 1.0 mM MgCl.sub.2) was added to two of three samples for each dilution. The samples were then incubated at 37 C. for 60 plus minutes. The absorbance of the supernatant was measured at 410 nm, and the values from appropriate control well(s) (e.g. total Nitrogen cavitated cell homogenate, nuclei and heavy mitochondria excluded) where substrate was not added were subtracted. The results were plotted and analysed.

(143) Selection Procedure: Competitive Bio-Panning Protocol

(144) Reaction Parameters

(145) 1.sup.st Selection Round

(146) n-CoDeR Lib2000 phage stock comprising 10.sup.10 genotype unique phagemid particles (Amp.sup.1) amplified to 210.sup.13 total pfu in 1.6 ml 2% milk-PBS (with Ca and Mg).

(147) Total reaction volume 2.5 ml

(148) Positive510.sup.7 Ramos B cell lymphoma cells

(149) NegativeJurkat T cell crude membranes derived from 210.sup.9 cells

(150) 2.sup.nd Selection Round

(151) 1.510.sup.12 phage eluted from previous selection round and then amplified, precipitated and re-suspended in 100 l 2% milk-PBS (with Ca and Mg).

(152) Total reaction volume 0.5 ml

(153) Positive510.sup.6 Ramos B cell lymphoma cells

(154) NegativeJurkat T cell crude membrane vesicles derived from 110.sup.9 cells

(155) 3rd Selection Round

(156) 110.sup.12 phage eluted and amplified from previous selection round re-suspended in 100 l 2% milk-PBS (with Ca and Mg).

(157) Total reaction volume 0.5 ml

(158) Positive510.sup.6 Ramos B cell lymphoma cells

(159) Negative Jurkat T cell plasma membrane vesicles derived from 110.sup.9 cells

(160) Method

(161) A 2 ml phage stock was pre-warmed at 3720 C. for 15 min and vortexed intermittently. The phage stock was centrifuged for 15 min at full speed in an eppendorf centrifuge. Where a precipitate had formed, the supernatant was transferred to a new eppendorf tube and resuspended in non-fat milk to a final concentration of 2%.

(162) Thaw control Jurkat cell plasma membrane preparations from 210.sup.9 cells (110.sup.9 cells biopanning rounds 2 and 3) on ice. (10 l was also saved for protein concentration determination.) The thawed plasma membrane preparations were resuspended by adding phage stock and by mixing with a pipette and subsequently incubated for 15 minutes on ice.

(163) 510.sup.8 (510.sup.6 cells biopanning rounds 2 and 3) Ramos cells were centrifuged at 1200 rpm, 6 min, 4 C.

(164) The supernatant was discarded and the Ramos cells resuspended in the milk-phage-negative cell membrane stock solution from biopanning round 2 and incubated at 10 C. and subjected to slow (end-over-end) rotation for 4 hours.

(165) The cell/cell membrane/phage incubate was transferred to a 15 ml Falcon tube containing 1 ml 100% (trypan blue stained) Ficoll at the bottom, and 9 ml overlaid 40% Ficoll-Paque Plus in 7% BSA/PBS (with Ca and Mg). Centrifuge at 1500 rpm for 10 min, 4 C. The tube was rotated 180 and centrifuged for a further 1 minute in order to dislodge cells from the tube wall.

(166) The interface containing whole Ramos cells and bound phage was carefully aspirated using a syringe and a higher gauge needle (e.g. Microlance 3-19GA11/2 1,140 TW PM). The needle is inserted just below the cell-containing interface with the bevelled end of the needle facing up. The cell layer is collected (approximately 150 l) and the needle pushed through the plastic of the tube opposite to the entrance hole. The contents of the syringe are expelled into a fresh tube, and washed twice by sucking up fresh PBS into the needle (still situated as piercing the tube). The harvest cell suspension was resuspended in 500 l of PBS-2% BSA and washing repeated, saving the supernatant for titration.

(167) Cells were resuspended in 1 ml PBS and transferred to a new 15 ml Eppendorf tube in which they were centrifuged at 1260 rpm for 10 min, 4 C. The supernatant was removed using a pipette saving the supernatant for titration.

(168) The phage were eluted from cells by addition of 150 l of 76 mM citric acid (pH2.5) in PBS followed by incubation at room temperature for 5 min. The mixture was neutralised by addition of 200 l of 1M Tris-HCl, pH 7.4. The cells were then centrifuged and the eluted phage saved.

(169) The cells were resuspended in 1 ml trypsin and transferred to a new tube and incubated for 10 min before inactivation with 40 l 1 mg/ml aprotinin. The cells were centrifuged, saving the supernatent for titration.

(170) Control Selection Protocol

(171) Control selection was performed using conventional negative pre-selection and subsequent positive selection using cell concentrations.

(172) Reaction Parameters

(173) Cells

(174) In all selection rounds

(175) Total reaction volume 4 ml

(176) Negative pre-selection 110.sup.8 Jurkat T cell cells

(177) Positive210.sup.7 Ramos B cell lymphoma cells

(178) 1.sup.st Selection Round

(179) n-CoDeR Lib2000 phage stock comprising 10.sup.10 genotype unique phagemid particles (Amp.sup.1) amplified to 210.sup.12 total pfu in 2.5 ml 2% milk-PBS (with Ca and Mg)

(180) 2.sup.nd Selection Round

(181) 2.610.sup.10 phage eluted from previous selection round and then amplified, precipitated and re-suspended in 2.5 ml 2% milk-PBS (with Ca and Mg)

(182) 3rd Selection Round

(183) 5.410.sup.10 phage eluted and amplified from previous selection round re-suspended in 2.5 ml 2% milk-PBS (with Ca and Mg)

(184) Method

(185) A 2.5 ml phage stock was pre-warmed at 37 C. for 15 min and vortexed intermittently. The phage stock was centrifuged for 15 min at full speed in an eppendorf centrifuge. Where a precipitate has formed, the supernatant was transferred to a new Eppendorf tube and subsequently re-suspended in non-fat milk to a final concentration of 2%.

(186) 110.sup.8 Jurkat T leukaemia cells were centrifuged at 1200 rpm, 6 min, 4 C. and resuspend in 1.5 ml 2% milk-PBS and subsequently by adding phage stock and by mixing with a pipette. They were then incubated for 24 hours at slow rotation, 4 C.

(187) The cell-phage incubation mixture was centrifuged at 300 g for 6 min, 4 C. The supernatant was discarded and washed three times in 15 ml PBS, re-centrifuged and the supernatant collected.

(188) 210.sup.7 Ramos cells were centrifuged at 300g, 6 min, 4 C. and resuspended in a collected phage mixture from negative preselection and incubated at 4 C. for 4 hours with slow agitation.

(189) The Ramos/phage incubation mixture was centrifuged and the supernatant discarded and the Ramos cells washed three times in 15 ml PBS.

(190) Phage were eluted from cells by addition of 150 l of 76 mM citric acid (pH2.5) in PBS followed by incubation at room temperature for 5 min. The mixture was neutralised by addition of 200 l of 1M Tris-HCl, pH 7.4. The cells were centrifuged and the elated phage saved.

(191) The cells were re-suspended in 1 ml trypsin, transferred to a new tube and incubated for 10 min. They were then inactivated with 40 l 1 mg/ml aprotinin, centrifuged and the supernatant saved for titration.

(192) Amplification on Large Plates Following Selection Rounds 1 and 2

(193) 1. 310 ml HB101F tubes were started (one for each selection to be performed+one for OD600 measurement) 2.5-3 h. before use. 50 l overnight culture to 10 ml LB containing 15 g/ml tet was used. OD is checked on one culture after approximately 2.5 h 2. The tubes were infected with half the eluted phage at OD.sub.600=0.5 3. The tubes were incubated for 30 minutes at 37 C. and 50 rpm, and for proper phenotyping an additional 30 min at 37 C., 200 rpm. 4. The bacteria were concentrated (10 ml) by centrifugation for 10 minutes at 2060g (3000 rpm Beckman GS-6). 5. They were then resuspended in part of the supernatant (approximately 3 ml) and spread on a large 500 cm.sup.2 plate (108 l/ml amp+15 g/ml tet+1% glu)/selection. 6. The plates were then incubated over night at 30 C. 7. The bacteria were collected from the plates by adding 5 ml of LB (100l/ml ampiciliin, 15 g/ml Tetracycline) per plate and the bacteria scraped off. The plate was tilted and the solution aspirated. 8. The plates were washed with an additional 3 ml LB medium as above and pooled with the first bacterial suspension a 50 ml Falcon tube. 9. The bacteria were concentrated by centrifugation for 10 minutes at 2100g/3000 rpm, Beckman GS6 at room temperature and resuspended in 1 ml of LB (+100 g/ml Amp and 15 g/ml Tet). 10. 500 l 50% glycerol was added to 1 ml (transferred from a larger volume as used in 5) and the glycerol stock frozen at 80 C. 11. 210 ml LB+amp (100 g/ml)+tet (15 g/ml) was infected with 2.5 l (5 l) of the glycerol stock of step 10, and grown until OD.sub.600=0.5. 12. 610.sup.9 PFU of R408 helper phage were added per ml culture and incubated for 30 minutes at 37 C. and 50 rpm. 13. IPTG solution was added to a final concentration of 100 M (i.e. 2 l from 0.5 M stock per 10 ml culture and incubated overnight at 25 C. (pili formation is inhibited at greater temperature) and 175 rpm.
Harvest and Precipitation of Amplified Phage Stocks 1. Bacteria were pelleted for 10 minutes at room temp. 2100g (3000 rpm, in Beckman GS-6) and the supernatant sterile filtered through 0.2 m sterile filter. 2. The appropriate tubes were pooled and the phage precipitated by adding volume of phage precipitation buffer and incubated for at least 4 hours at 4 C. (and can be incubated for weeks). 3. The tubes were then centrifuged for 30 minutes at 4 C. and 13000g. 4. The pellet was resuspended completely in 100 l PBS over night at 4 C.
Titration of Phage
Materials: Overnight culture of E. coli HB 101F v-bottom microtitre plate 1PBS (sterile) LB-medium LA+amp (100 g/ml)+tet (15 g/ml)+1% glucose plates 1. 10 ml LB+tet (15 g/ml) was inoculated with 100 l of an over-night culture of E. coli HB101 F and incubated at 37 C., 175 rpm, until OD.sub.600=0.5 (approx. 2-2.5 h). 2. For amplified phage stocks, the phage were diluted 10.sup.5 times in PBS in a v-bottom microtitre plate before infection. Eluted phage are used undiluted for infection (see below). 3. When the E. coli cells reached OD600=0.5, 100 l of the culture was transferred to every second well of the first column of wells on a v-bottom microtitre plate. The three following wells in the row were filled with 90 l LB-mediurn. 4. 10 l of the phage stocks were transferred (diluted or undiluted as above) to 100 l bacteria in the microtitre plate and incubated for 30 min at 37 C., 50 rpm. 5. The infected bacteria were diluted by sequentially transferring 10 l from each well to the next well filled with 90 l LB. 8. 10 l from each well was spotted onto a dry LA+amp+tet+1% glu plate using a multi-channel pipette with every second tip removed, i.e. with only four tips, 45 spots were spotted onto on each LA plate and allowed to dry before inverting the plates for incubation at 37 C. 9. The colonies were counted and the litre (CFU/ml) calculated, (CFUdilution100).
Titrations Performed

(194) TABLE-US-00004 TABLE IV Dilutions After Start pool 10.sup.7 10.sup.8 10.sup.9 10.sup.10 10.sup.11 Selection HB101F BnonT-C.1 10.sup.2 10.sup.3 10.sup.4 10.sup.5 10.sup.6 Round 1 eluate After HB101F BnonT-C.1 10.sup.7 10.sup.8 10.sup.9 10.sup.10 10.sup.11 Selection amplified Round 2 HB101F BnonT-C.2 10.sup.5 10.sup.6 10.sup.7 10.sup.8 10.sup.9 After HB101F BnonT-C.2 10.sup.7 10.sup.8 10.sup.9 10.sup.10 10.sup.11 Selection amplified Round 3 HB101F BnonT-C.3 10.sup.5 10.sup.6 10.sup.7 10.sup.8 10.sup.9
Amplification on Plates for Glycerol Stocks, and Over Night Culture for Minipreps (Following Selection Round 3).
Reaction Materials

(195) 310 ml HB101F with OF 0.5 are needed (10 ml for each target and 10 ml for titration)

(196) 3 large 500 cm.sup.2 plates are needed (amp+tet+glu).

(197) Method

(198) For each selection: 1. Half of each eluate from selection 3 is used to infect 10 ml log phase bacteria (OD.sub.600=0.5) which are then incubated for 30 min. at 37 C., 50 rpm and grown for 30 min at 37 C., 175 rpm. 10 ml warm media containing 200 g/ml Ampicillin is added and divide in 2 parts of 10 ml each. 2. The bacteria were concentrated (10 ml) by centrifugation for 10 minutes at 2100g/3000 rpm, Beckman GS-6 at room temperature and resuspended in a small volume to be spread on one 500 cm.sup.2 plate (100 g/ml amp+15 g/ml tet+1% glu) and grown over night at 30 C. 3. Miniprep: 10 ml were spun down and resuspended in 6 ml LB containing 0.1% glucose and 100 g/ml amp and Grow over night at 30 C. 4. The bacteria were collected from the plates by adding 3-5 ml of LB (100 g/ml ampicillin, 15 g/ml Tetracycline) per plate and scrape of bacteria. Tilt the plate and aspirate the solution. 5. The bacteria were concentrated by centrifugation for 10 minutes at 2100g/3000 rpm, Beckman GS6 at room temperature and resuspended in 1 ml of LB (+Amp/Tet as above). Add 500 l 50 % glycerol and the glycerol stock frozen at 80 C. 6. Minipreps were prepared from 3 ml of culture according to protocol from the kit manufacturer (BioRad).
Conversion to scFv Format after Selection 3

(199) The pMil vector was digested with EagI, an enzyme that cuts on both sides of gene III. The vector was then re-litigated followed by a killer cut using EcoRI. This enzyme has a site in gene III and thus will destroy non-converted vectors.

(200) A digestion mix was prepared using

(201) TABLE-US-00005 Phagemid 6.25 l (0.21-0.66 g) NEBuffer 3 0.75 l EagI (10 U/l) 0.5 l MQ-H.sub.2O Added to 7.5 l and incubated at 37 C. for 2 h with heat inactivation for 20 min at 65 C.

(202) Ligation of the vector is performed with

(203) TABLE-US-00006 EagI digested phagemid 3 l MQ-H.sub.2O 20 l Incubated for 5 min at 45 C. and put on ice. 5 ligase buffer 6 l T4 DNA ligase (1 U/l) 1 l 30 l Incubated for 2 hours at room temperature and heat inactivated at 65 C. for 10 minutes

(204) The killer cut is performed with

(205) TABLE-US-00007 Ligated material 30 l 10 REACT3 3.6 l EcoRI (10 U/l) 2 l 1M NaCl 0.4 l MQ-H.sub.2O 4 l 40 l incubated for 90 min at 37 C. and heat inactivated for 10 minutes at 65 C.

(206) The ligate is stored at 4 until used.

(207) Transformation

(208) The ligate produced above was transformed into E. coli TOP10, see table below.

(209) One tube (100 l/tube) of chemically competent TOP10 cells were thawed and incubated on ice for 10 min. 10 ng ligate were added per tube (3.2 l ligates made above) and incubate on ice for 30 min.

(210) The tubes were then incubated at 42 C. (water bath) for 90 s, and further incubated on ice for 1-2 min.

(211) 900 l LB was added to each tube and incubated at 37 C., 1 h, 200 rpm.

(212) The content of each tube was spread on one large LA-Amp (100 g/ml)-Glu (1%) plate, and incubated overnight at 37 C.

(213) Colony Picking

(214) A total of 960 clones were picked from each of the two selections from the large LB plates to 1096 well microtitre plates using a Genetix Q-bot colony picking system. 1) LB media was mixed with appropriate antibiotics and glucose 1% and Greiner Flat 96 well 655501 plates filled with media, 150 l/well. 2) Colonies were picked with the Q-Bot according to the protocol, and the plates incubated at 37 C. over night without shaking.
Expression of Clones in 96 Well Format 1) The expression plate (Greiner Round 96 well 650101) was filled with media 100 l/well including appropriate antibiotics and inoculated with 5 l/well from the master plate and the expression plate incubated at 37 C., 600 rpm for 3.5 h. 2) The production of scFv was induce with 25 l/ well of media including appropriate antibiotics and 2.5 mM IPTG. 3) The expression plate was incubated for 10 h at 37 C., 600 rpm. 4) 80 l of 1M PBS was added to the filter plate and filter added to the stock plate. 5) 80 l was transferred from the expression plate to the filter plate and filter to the stock plate. The Stock plate was stored at +4 C.
Screening of Clones in 96 Well Format

(215) Clones were screened in a macro confocal FMAT 8100 HTS system (Applied Biosciences, Foster City, Calif., USA). Gates were set using B and T cell specific monoclonal antibodies (anti-CD 19 and anti-CD3 respectively) so that B cell positive signal was obtained only by staining with anti-CD19 primary antibody, and T cell positive signal only with anti-CD3 primary antibody. (FL1<10666, FL2<2026, Colour:0.17-0.425, Size 10.0-66.66).

(216) Materials

(217) CD45 0.5 mg/ml (BD Biosciences cat no 555480) CD19 0.5 mg/ml (BD Biosciences cat no 555410) CD3 0.5 mg/ml (BD Biosciences cat no 555330) Purified clone 31 0.154 mg/ml 010515 KMN -HIS (R&D Systems MAB050, 8.43 mg/ml) mouse-Cy5 (Amersham Bioscience PA45002) PBS (Gibco, Cat no 14040-091) Hepes Buffer (Gibco, Cat no 15630-056) Buffer: PBS+10% Hepes buffer Cell media: RPMI 1640 with 10% FCS, 10 mM Hepes, 1 mM Sodium Pyruvate LB media
Method 1) 10,000 cells/well diluted in cell media, were added to 40 l/well incubate grow on at 37 C., 5% CO.sub.2. 2) 10 l/well of scFv were added to one plate with B cells and one plate with T cells. Also added were -CD45 to column 2, -CD19 to column 3 and CD3 to column 4, and the plates incubated for 1 hour. 3) -mouse-Cy5 diluted in buffer were added, 20 l/well (dilution 1:1000, final dilution, working dilution 1:286) to the control plates. mouse cy5+His diluted in buffer (dilution 1:1000 and 1:5000 final dilution, 1:286 and 1:1429 working dilution) were also added. 4) The plates were read on the FMAT 8100 HTS systems after 4 h. The reading repeated after an over night incubation.

EXAMPLE 11

Pharmaceutical Compositions

(218) A further example of the invention provides a pharmaceutical composition comprising an anti-ligand (the active ingredient) isolated according to the method of the invention.

(219) The aforementioned compounds of the invention or a formulation thereof may be administered by any conventional method including oral and parenteral (e.g. subcutaneous or intramuscular) injection. The treatment may consist of a single dose or a plurality of doses over a period of time.

(220) Whilst it is possible for a compound of the invention to be administered alone, it is preferable to present it as a pharmaceutical formulation, together with one or more acceptable carriers. The carrier(s) must be acceptable in the sense of being compatible with the compound of the invention and not deleterious to the recipients thereof. Typically, the carriers will be water or saline which will be sterile and pyrogen free.

EXAMPLE 12

Screening Without Cell Culture

(221) The present invention allows for biomarker and target discovery directly at the protein level by use of e.g. highly diversified molecular libraries. The present prophetic example demonstrates its applicability for selective isolation of binders with specificity to antigens upregulated or unique in one plasma sample (e.g. from a cancer patient) compared to another (from a healthy control).

(222) A highly diversified molecular library containing 10.sup.11 genotype unique members at a binder copy number of approximately twenty (effective number of binders displaying binding ability, non-displaying examples are removed by means of affinity purification) is precipitated by polyethyleneglycol treatment and re-suspension in 20 l 10% milk-PBS solution. Isolated binders from selection using this library have nanomolar affinity (Kd=110.sup.9) for the relevant antigen.

(223) Whole plasma (albumin and some other high abundance large protein that are not suspected to differ between the two patients are removed by size exclusion chromatography) from a cancer patient and a healthy control are collected and treated as follows:

(224) The cancer patient plasma sample is split and biotinylated at three different amino acid positions, in order to minimise the loss of relevant antigen epitopes being destructed by the chemical biotinylation process.

(225) The control plasma is subjected to a mock treatment (treatment without addition of biotin). Control plasma is then concentrated forty times by appropriate treatment, and is subsequently mixed with the plasma sample from the cancer patient in equal volumes (total volume=1.0 ml).

(226) The mixture produced is mixed with the blocked molecular library above, and the produced plasma/binder mixture is incubated over night at 4 C. at slow end-over-end rotation.

(227) Following selection biopanning, capture beads e.g. streptavidin coated magnetic particles (streptavidin coated MACS from Milenyi, USA; bead diameter=50 nm) are added and used to isolate biotinylated proteins from the cancer patient plasma sample and antigen bound molecular library binders (Siegel et al., 1997).

(228) The magnetic beads are added at a molar ratio sufficient to capture all biotinylated antigen added (equivalent of total cancer patient biotinylated plasma protein concentration).

(229) The mixture is transferred to a magnetic separation column pre-treated as described by Siegel, L. et al., J Immunol Methods, 206 (1997), pp73-85: 1. Before addition of selection suspension, the MACS column is loaded with 2% BSA-PBA (BPBS) to fill the lower part (that will later be outside the magnetic field), and so that subsequently added selection mix will distribute evenly through the magnetic part of the column without clogging. A 30-gauge inch needle is mounted at the columns outlet port to restrict the flow rate to 10 l/min. 2. Once loaded, the column is placed inside the magnet for 2 min. 3. A total of 3 washes with BPBS are performed for each selection round, the first of which is performed with the needle attached to the column outlet port, and sub-sequent washings performed without a flow-restricting needle (flow-rate 200 l/min). 4. Following the last wash, the column is removed from the magnet and the bead-coated/phage coated/biotinylated antigen are flushed off the column with 500 l PBS using the plunger from a 5-cc syringe (Beckton-Dickinson, Franklin Lakes, N.J., USA). 5. Elution is achieved by acid, base or otherwise suitable means (e.g. enzymatic cleavage if the molecular library members contain a specific enzyme cleavage site). 6. The beads are centrifuged, and eluted binders are saved and titrated.

(230) Using the above experimental set-up, one may expect to isolate binders with specificity for antigens 10-fold or more upregulated (or uniquely present) in the patient plasma sample compared to the control plasma sample and present at sub-nanomolar levels (or greater) in the cancer patient sample, whilst eliminating binders with specificity for antigens present at equal abundance in the original cancer patient and control patient samples.

(231) Table V strews expected relative numbers (theoretical, assuming no dissociation of nanomolar binders during washing) of binders with specificity for different categories of antigen.

(232) TABLE-US-00008 TABLE V Kd = 1 10.sup.9M Healthy control antigen Binders Cancer patient plasma plasma antigen specific isolated Binder antigen concentration concentration binder copy following Category (moles/litre) (moles/litre) number input biopanning Positive cell 3.3 10.sup.8 0.00E+00 2.00E+0 19 restricted antigen Positive cell 3.3 10.sup.8 3.3 10.sup.9 2.00E+01 4 enriched antigen Positive cell 3.3 10.sup.8 3.3 10.sup.10 2.00E+01 14 enriched antigen Positive cell 3.3 10.sup.9 3.3 10.sup.10 2.00E+01 4 enriched antigen Positive cell 3.3 10.sup.10 3.3 10.sup.11 2.00E+01 2 enriched antigen Positive cell 1 10.sup.10 0 2.00E+01 1 enriched antigen Common antigen 3.3 10.sup.6 3.3 10.sup.6 2.00E+01 0 (highly expressed) Common antigen 3.3 10.sup.8 3.3 10.sup.8 2.00E+01 0 (moderately expressed)

(233) The molecular library may be substituted for by synthetically produced chemicals. It would still be necessary to approximate the highest probable affinity of target chemicals in the chemical library to be isolated and the number/concentration of each modified chemical compound added.

EXAMPLE 13

Using the Equations to Determine the Suitability of a Known Molecular Library in Performing the Invention

(234) The equations of the invention can be used to evaluate molecular libraries of known parameters (including of known amounts) for their suitability to isolate anti-ligands with a given specificity. E.g. use of the equations can determine the amount of subtractor and/or target ligand required to isolate the requisite binders and may be compared to the amounts known in a provided sample, in order to determine if a sufficient amount of ligand is present.

EXAMPLE 14

Performing the Invention by Automatic Methods

(235) Screening of anti-ligands using the method of the invention may be performed using varying levels of automation. The determination of the amount of first subtractor and second target ligand can be performed using a computer program designed to run the equations derived from the law of mass action.

(236) Furthermore the method of screening may be further automated if necessary for high throughput screening. In this situation, the automatic determination of the amount of first and second target ligands, is followed by automatically providing the determined amounts of first and second target ligands and automatically exposing the ligands to the anti-ligand.

(237) An outline for a computer software program utilising the invention to automate the binder selection process of the present invention, Is described below:

(238) The user will be asked to choose a selection program. In simple mode such programs would, include: A. Positive to negative inter-population ligand abundance dependent analysis and positive intra-population ligand abundance independent isolation of binders (higher stringency-higher antigen concentrations needed). This program is ideal for biomarker or target discovery. (e.g., FIG. 5) B. Positive to negative inter-population, ligand abundance dependent analysis and positive population infra-population ligand abundance dependent isolation of binders (lower stringency-lower antigen concentrations needed) (e.g., FIG. 2). This program would have different sub-programs, e.g. transfected-cell selection, wild-type cell selection, or plasma selection. Depending on which type of selection is chosen the software program will estimate the relative frequencies of target ligands present at different absolute and relative numbers in the positive and negative ligand populations. The user may also choose to enter these numbers manually.

(239) The user will be asked to specify which category of binders is of interest. The user may choose to perform selection such that only binders with specificity for positive population unique ligands are isolated, or choose to include binders with specificity for ligands present at higher concentration in the positive ligand population compared to the negative ligand population. The user will also provide information about the binder library: 1. Diversity of library (number of unique binders). This value is used to set higher threshold for binder affinity. 2. Library binder copy numberthe multiplicity of each unique binder member 3. Selection reaction volume (liters) 4. Threshold for binder affinity (Kd (M))This value is the lowest affinity of interest of any binder for its target ligand. The upper threshold for this value is determined from 1. 5. The expected number of times each binder will be multiplied following a selection round (The amplification factor). The amplification factor is calculated as the number of phage following amplification of binders isolated from a given selection round divided by the number of binders isolated from the same selection round. The expected amplification factor is replaced by the actual amplification factor calculated as selection rounds are performed so that each selection round is optimised. 6. Lowest accepted number of unique binders isolated following each selection round. This number could be anything from less than 1 to the input binder copy number. 7. Availability of subtractor ligand constructs (the highest number of subtractor constructs that are available and may be used in the selection process) 8. Ligand detection level. The approximate concentration of the lowest, abundant target ligand of interest in the positive ligand population. Binders specific for target ligand present at lower concentrations than that specified here may not be expected to be isolated. 9. Desired frequency of binders with specificity for the most rare or weakly upregulated ligands isolated following the final round of selection.

(240) Given the above parameters, the computer program will calculate the number of selection rounds necessary for binders with sought specificity and affinity to be isolated at desired frequency (based on entered parameter values), and the number of positive and negative ligand constructs to use in the selection rounds respectively. All or none of the parameters entered for a given selection may be saved and used for subsequent selections.

(241) Furthermore, the output concerning the quantities of any material and the number of selection rounds can be linked through a software program to an automated experimental set-up where by the user inputting data allows the automatic calculation and implementation of the requisite experimental materials.

REFERENCES

(242) Barbas C F, Kang A s, Lerner R A and Benkovic S J, (1991), Assembly of combinatorial antibody libraries on phage surfaces: the gene III site. Proc. Natl. Acad. Sci. USA 88, 7978-7982. Carlsson, R., Fischer, H. and Sjogren, H. O. (1988) Binding of staphylococcal enterotoxin A to accessory cells is a requirement for its ability to activate human T cells, J Immunol 140, 2484. Chiswell D J and McCafferty J, (1992) Phage antibodies: will new coliclonal antibodies replace monoclonal antibodies? Trends Biotechnol. 10, 80-84. Clackson T, Hoogenboom H R, Griffiths A D and Winter G, (1991), Making antibody fragments using phage display libraries. Nature 352, 624-628. de Kruif J., Terstappen, L., Boel, E. and Logtenberg, T. (1995) Rapid selection of cell subpopulation-specific human monoclonal antibodies from a synthetic phage antibody library. Proc Natl Acad Sci USA 92, 3938. Dower, W. J., CWIRLA, S. B, and N. V., A. T, (1991) Recombinant Library Screening Methods. In: World Intellectual Property Organization. SMITH, W. M., United States. Felici F, Luzzago A, Manaci P, Nicosia A, Sollazzo M and Traboni C, (1995), Peptide and protein display on the surface of filamentous bacteriophage. Biotechnol. Annual Rev. 1, 149-183. Francisco, J. A., Stathopoulos, C., Warren, R.A. Kilburn, D. G. and Georgiou, G. (1993) Specific adhesion and hydrolysis of cellulose by intact Escherichia coli expressing surface anchored cellulase or cellulose binding domains. Biotechnology (N Y) 11, 491. Gao, C., Lin, C. H., Lo, C. H., Mao, S., Wirsching, P., Lerner, R. A. and Janda, K. D., (1997) Making chemistry selectable by linking it to infectivity. Proc Natl Acad Sci USA 94, 11777. Hanes, J. and Pluckthun, A. (1997) In vitro selection and evolution of functional proteins by using ribosome display. Proc Natl Acad Sci U S A 94, 4937. He, M. and Taussig, M. J. (1997) Antibody-ribosome-mRNA (ARM) complexes as efficient selection particles for in vitro display and evolution of antibody combining sites. Nucleic Acids Res 25, 5132. Hoogenboom H R. and Winter G. (1992), By-passing immunisation. Human antibodies from synthetic repertoires of germline VH gene segments rearranged in vitro. J. Mol. Biol. 227, 381-388). Hoogenboom H R, deBruine A P, Button S E, Hoet R M, Arends J W and Rooves R C, (1998), Antibody phage display technology and its application. Immunotechnology 4(1), 1-20. Huls, G. A., Heijnen, I. A., Cuomo, M. E., Koningsberger, J. C., Wiegman, L., Boel, E., van der Vuurst de Vries, A. R., Loyson, S. A., Helfrich, W., van Berge Henegouwen, G. P., van Meijer, M., de Kruif, J. and Logtenberg, T. (1999) A recombinant, fully human monoclonal antibody with antitumor activity constructed from phage-displayed antibody fragments. Nat Biotechnology 17, 276. Jacobsson, K. and Frykberg, L. (1995) Cloning of ligand-binding domains of bacterial receptors by phage display, Biotechniques 18, 878. Katz B A, (1997), Structural and mechanistic determinants of affinity aid specificity of ligands discovered or engineered by peptide display. Annual Rev. Biophys. Biomol. Struct. 26, 27-45. Kay B K and Paul J I, (1996), High-throughput screening strategies to identify inhibitors of protein-protein interactions. Mol. Divers. 1, 139-140). Koide, A., Bailey, C. W., Huang, X. and Koide, S. (1998) The fibronectin type III domain as a scaffold for novel binding proteins. J Mol Biol 284, 1141. Mandecki, W., Chen, Y. C. J. and Gribalde N (1995) A mathematical model fro biopanning (affinity selection) using peptide libraries on filamentous phage, J. theor. Biol. 176, pp 523-530. Markland, W., Roberts, B. L., Saxena, M. J., Guterman, S. K. and Ladner, R. C. (1991) Design, construction, and function of a multicopy display vector using fusions to the major cost protein of bacteriophage M13. Gene 199, 13. Marks J D, Hoogenboom H R, Bonnert T P, McCafferty J, Griffiths A D and Winter G, (1991), By-passing immunisation, Human antibodies from V-gene libraries displayed on phage. J. Mol. Biol. 222, 581-597. Mattheakis, L. C., Bhatt, R. R. and Dower, W. J. (1994) An in vitro polysome display system for identifying ligands from very large peptide libraries. Proc Natl Acad Sci USA 91, 9022. McCafferty, J., Griffiths, A. D., Winter, G. and Chiswell, D. J. (1990) Phage antibodies: filamentous phage displaying antibody variable domains. Nature 348, 552. Siegel, D. L., Chang, T. Y., Russell, S. L. and Bunya, V. Y. (1997) Isolation of cell surface-specific human monoclone antibodies using phage display and magnetically-activated cell sorting: applications in immunohematology. J Immunol Methods 206, 73. Smith, G. P. (1985) Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science 228, 1315. Smith, G. P. and Scott, J. K. (1993) Libraries of peptides and proteins displayed on filamentous phage. Methods Enzymol 217, 228. Stahl, S., Sjolander, A., Nygren, P. A., Berzins, K., Perlmann, P. and Uhlen, M. (1989) A dual expression system for the generation, analysis and purification of antibodies to a repeated sequence of the Plasmodium falciparum antigen Pfl55/RESA. J Immunol Methods 124, 43. Stausbl-Grn, B., Wind, T., Kjaer, S., Kahns, L., Hansen, N., Kristensen, P., and Clark, B. (1996) A model phage display subtraction method with potential for analysis of differential gene expression, Febs Letters, 391, pp 71-75. Stausbl-Grn, B., Jensen, K. B., Jensen, K. H., Jensen, M. and Clark, B. (2001) De novo identification of cell-type specific antibody-antigen pairs by phage display subtraction, Eur. J. Biochem., 268, 3099-3107. Weng, S., Gu, K., Hammond, P. W., Lohse, P., Rise, C., Wagner, R. W., Wright, M. C. and Kuimelis, R. G. (2002) Generating addressable protein microarrays with PROfusion covalent mRNA-protein fusion technology. Proteomics 2, 48. Williams, B. R. and Sharon, J. (2002) Polyclonal anti-colorectal cancer Fab phage display library selected in one round using density gradient centrifugation to separate antigen-bound and free phage. Immunol Lett 81, 141. Winter, G., Griffiths, A. D., Hawkins, R. E. and Hoogenboom, B. R. (1994) Making antibodies by phage display technology. Anna Rev Immunol 12, 433. Winter and McCafferfy (1996) Phage display of peptides and proteins: a laboratory manual Ed Kay, Academic Press, Inc ISBN 0-12-402380-0.