Colin H. Self

Light activatable antibodies: Models for remotely activatable proteins

The ability to activate biological macromolecules remotely, at specific locations and times, will allow the manipulation of a wide range of cellular activities and give rise to many practical applications. Interest has been shown in the theoretical possibility of accomplishing this by means of photochemical approaches1. Photochemical changes of the guest–binding cavity of cyclodextrins has been suggested2; however, these changes require organic solvent. What is needed is a widely and readily applicable method allowing activation under physiological conditions. We have developed such a method. This is based on our demonstration that relatively large amounts of the a–methyl substituted 2–nitrobenzyl alcohol, namely, 1–(2–nitrophenyl)ethanol (NPE) can be coupled to proteins using diphosgene3,4. Previous work involved “caging” of small molecules such as ATP (ref. 5–9) and blocking amino acids10 in peptide synthesis11,12 with 2–nitrobenzyl compounds. For large molecules, site–specific reversible inactivation of T4–lysozyme has been reported following introduction of an aspartyl β–nibenzyl ester into its active site by mutagenesis13. In contrast, the present simple procedure allows an existing protein to be deactivated and then, when and where required, reactivated by exposure to ultraviolet–A (UV–A) light. We have employed antibodies as models for both receptors and ligands and have successfully modulated: antibody binding sites for antigen; antigen binding sites for antibody, and antibody Fc binding sites for Protein A.

A highly specific and sensitive enzyme-immunometric assay for calcitonin gene-related peptide based on enzyme amplification

Calcitonin gene-related peptide (CGRP) is an important member of the peptide family encoded by the calcitonin gene. It has been found to be a potent vasodilator in man and a major circulating gene product (Girgis et al., 1985). The present study reports the development of a sensitive and rapid two-site immunoassay for CGRP based on enzyme amplification (Self, 1985). The assay has been easy to construct, taking advantage of available antisera raised for other purposes. Nevertheless it has been found to be clearly superior to our previous radioimmunoassay in terms of sensitivity, specificity, speed and convenience.

Light activation of anti-CD3 in vivo reduces the growth of an aggressive ovarian carcinoma.

In the preceding article we describe the controlled photoactivation of human T-cells by their illumination in the presence of the previously inactivated anti-human CD3 antibodies, OKT3 or UCHT1. The photoreversible inhibition of the anti-human CD3 antibodies uses a coating of photocleavable 2-nitrophenylethanol (NPE) groups to block the antibody’s activity. The NPE groups could be removed by irradiation with UVA light, leaving the antibodies free to bind to, and activate, human Tcells at any given time and place. This provided a means to target human T-cells with a much higher degree of specificity to tumours, whilst minimising side effects in other non-illuminated areas, and was the first key step in our ultimate goal: the preparation of highly specific photoactivatable bispecific tumour-targeting antibody conjugates. The next steps prior to the construction of the cancer-targeting conjugate were to examine 1) if an NPE-inactivated anti-Tcell antibody could be reactivated in vivo and in vitro, and 2) whether such an activation would have any effect on the growth of a tumour. We decided to switch to a C57BL6 murine system to carry out this work. Low levels of the hamster antimurine CD3 monoclonal antibody 145-2C11 (a hamster equivalent of OKT3), had been shown to prevent malignant tumour progression, and we had frozen stocks of several virulent tumours which grew in C57BL6 mice. As 145-2C11 targets C57BL6 T-cells, we could obtain all the components to examine tumour growth in a syngeneic C57BL6 model system. The 1452C11 hybridoma was obtained, and the antibody was coated with NPE in a similar manner to that used to inhibit OKT3 and UCHT1. A progressive ovarian tumour, M5076, was selected for study, and the effects of the addition of 145-2C11, NPEcoated 145-2C11, in vitro UV-irradiated NPE–145-2C11, and in vivo UV-irradiated NPE–145-2C11 antibody on the growth of the tumour was then examined. The 145-2C11 monoclonal antibody was quite susceptible to coating, with less NPE–COCl required to fully inhibit its activity (relative to the anti-human CD3 antibodies). This caused the antibody to be prone to precipitation during dialysis to remove the excess NPE, and the large majority (>95 %) spun out on centrifugation at 13 000 rpm. When resuspended in isotonic saline, this NPE-coated 145-2C11 antibody could not bind to the CD3-expressing murine lymphoma cell line, EL4. However, after illumination of this NPE-coated antibody with UV light for 10 min in the presence of the EL4 cells, considerable binding of the antibody (70 % of that given by uncoated antibody) could be detected (Table 1). This was an extremely important result demonstrating that we could regulate the anti-murine CD3 antibody in a similar manner to the anti-human CD3 antibodies, OKT3 and UCHT1.

Enhanced specificity for small molecules in a convenient format which removes a limitation of competitive immunoassay

Anti-complex immunoassay systems for small molecules permit the exquisite specificity achievable with monoclonal antibodies to be expressed to an extent which is not possible with competitive format immunoassay. While our previously reported anti-complex system is superior to competitive systems in terms of sensitivity, precision and specificity we have found that this specificity may be enhanced dramatically by simply interposing a wash step between the addition of sample and that of the labelled anti-complex antibody. When such a wash step was attempted with the competitive format system, after addition of sample but before addition of the labelled component, assay performance was degraded to the extent of making it unusable. We suggest, therefore, that the inherent flexibility of the anti-complex approach for small molecule assay creates an opportunity for remarkable enhancement of the functional specificity of primary antibody.

Reduction of non-specific binding in immunoassays requiring long incubations

Despite the studies so far about the non-specific binding of antibody molecule to the plastic of solid phase in enzyme-linked immunoassays, background binding in microwell Elisa continues to be a troublesome problem.Non-specific immunoglobulin from an undiluted serum sample can adhere to the surface of a ‘blocked’ plate to result in a maximal signal in an antigen capture assay for specific antibody to render analysis virtually impossible in undiluted serum when using labelled anti-species antibodies. Yet it is desirable in many circumstances that the maximum sensitivity achievable by the simple expedient of using a concentrated sample (undiluted serum) be exploited, for example in the analysis of antibodies to HIV in the interest of earlier diagnosis. To circumvent this problem we have developed an alternative strategy in which a biotinylated capture reagent is preincubated with the serum sample for the necessary time after which the biotinylated ligand/antibody complex is itself rapidly captured in streptavidin-coated wells at 4°C, with subsequent detection with labelled anti-species immunoglobulin. This manoeuvre enables the capture ligand to be incubated with undiluted serum sample for long time periods resulting in improved specificity of detection. By this means we describe a general method to improve the specificity of serum antibody immunoassays which will be expected to produce the benefit of more rapid diagnosis by signalling antibody production earlier in the abnormal state. Furthermore, our new method could be used to reduce non-specific binding in other immunological assays such as antibody arrays to which much attention has recently been paid.

Light-activated antibodies in the fight against primary and metastatic cancer.

The very cytotoxic potency of therapeutic antibodies used in the fight against cancer makes their specific tumour targeting of crucial importance. Unfortunately, in practice, this is often not achieved and can lead to dangerous side-effects. A way of greatly reducing such side-effects is to make the antibodies region-specific to the areas bearing tumour. This can now be achieved by rendering them light dependent so they are only active where illuminated. There are many ways of employing such light-enhanced targeting in very many locations within the body. When it is applied to direct killer T-cells to ovarian primary tumours, not only is primary tumour growth markedly reduced but also a dramatic reduction of metastatic growth is observed in the liver.

Light‐Directed Activation of Human T‐Cells

Photocleavable “caged” biomolecules have been successfully used as photoswitches and phototriggers as reported in many biological studies, but relatively few of these reports have described the effective caging and uncaging of proteins under physiological conditions. We have developed a procedure that uses a coating of photocleavable 2-nitrophenylethanol (NPE) groups to reversibly inhibit antibodies. The antibodies can be reactivated, when and where required, by irradiation with UVA light. Reactivation occurs even when the inactivated antibodies are irradiated through plastic in the presence of cells. This led us to suggest that this procedure could be used to greatly increase the specificity of therapeutic antibodies, 5] and we recently described the construction of a photoactivatable cancer-targeting bispecific conjugate. We proposed that if only the cytotoxic end of such a conjugate were to be reversibly inactivated, then the tumour-specific end would remain free to bind to its target tumour cells without damage to normal tissues also targeted through specific and nonspecific cross-reactions. Localised illumination of the tumour-targeted conjugate reactivates the cytotoxic end, maximising tumour destruction whilst minimising damage to healthy tissue. Such a technique could also be used to improve the targeting of a patient’s immune response to a tumour; this has been suggested as an elegant alternative to the use of toxins or enzymes. After early studies suggested that a low-level activation of Tcells in serum could prevent malignant tumour progression, bispecific antibodies were designed to directly target cytotoxic T-cells to tumours. In theory, one part of the antibody reacts with a specific tumour antigen and binds to the tumour cell surface whilst the other part of the antibody reacts with a Tcell marker (normally CD3), hence targeting the T-cell to the tumour cell (Figure 1). In practice this procedure suffers from two major drawbacks : Firstly, it has proved to be very challenging to obtain the exquisite specificity required to clinically differentiate between tumour and normal cells, perhaps not surprising when the large ratio of normal tissue to tumour tissue in most patients is taken into consideration. This limits the degree of specific localisation that can be achieved against certain tumours. To date relatively few antibodies have been licensed for use against solid tumours. Secondly, the introduction of anti-T-cell-based bispecific constructs results in them being bound by peripheral T-cells before the bispecific antibody reaches its tumour target. This both impedes the bispecific antibody and activates T-cells peripherally, leading to Tcell depletion and unwanted cytokine storms. Both of these challenges could be effectively overcome if the anti-CD3 part of the bispecific antibody could be reversibly inactivated. The antitumour portion of the antibody would remain free to circulate and bind to tumour cells, but the anti-CD3 portion would not be able to bind, activate, and remove peripheral Tcells from the patient’s circulation. Nonspecific cross-reactions or specific unwanted binding of the antitumour antibody would become irrelevant, as the anti-CD3 portion of the antibody would be inactive until deliberately reactivated in the area of the tumour (Figure 1). An added bonus would be that higher doses of conjugate could be administered, allowing more conjugate to target the tumour. This report demonstrates the production of the most important component necessary to achieve this goal : photoreversibly inactivated anti-human CD3 antibodies. A coating of NPE groups is used to block the activity of OKT3 or UCHT1 (two anti-human CD3 antibodies). On illumination with UVA light, the NPE groups cleave, leaving the antibodies free to bind to, and activate, the human H9 T-cell line. A coating of NPE groups was used to inhibit the biological activity of OKT3. 30 mL of NPE–COCl was used to coat 1 mL of antibody. The final yield of soluble inactivated OKT3 was routinely 20% (0.2 mgmL ) with approximately 50 NPE residues present on each antibody molecule (as determined by the increase in OD280 of the antibody). The inhibition and reactivation of OKT3 binding to the H9 T-cell line was then investigated by flow cytometry. As UV irradiation is needed to remove NPE from the coated OKT3 samples, and it is known to damage some biological molecules, it was first important to demonstrate that UV irradiation does not damage uncoated OKT3. Figure 2 shows Figure 1. How a bispecific antibody conjugate links a T-cell to a tumour. This diagram shows the advantages of reversibly inactivating the anti-CD3 (OKT3 or UCHT1) portion of a cancer-targeting conjugate with NPE. The conjugate is reactivated only where it is bound to tumour by irradiation with UV light; TA= tumour antigen. Throughout these studies we used a human T-cell clone (H9), which expresses CD3 on its cell surface.

A general method for the complete deglycosylation of a wide variety of serum glycoproteins using peptide-N-glycosidase-F

Many indirect serum studies show changes in protein glycosylation in disease, but further progress will require direct investigation of oligosaccharide composition. Current methods of deglycosylation using PNGase-F often result in incomplete removal of oligosaccharides. This is an unsatisfactory situation because only small quantities of material are often available in clinical studies, glycosylation changes may occur in only a small proportion of the molecules and quantification of the released oligosaccharides may be unreliable. The ability of PNGase-F to deglycosylate haptoglobin (Hp) under different conditions has been investigated. Oligosaccharides were completely removed from 50μg of Hp by treatment for 24 h with PNGase-F in 50 mmol/l ammonium formate buffer, pH 8.6, in combination with sodium dodecyl sulphate, mercaptoethanol and Nonidet P40. This modified procedure was equally effective at removing oligosaccharides from other serum glycoproteins (α1 glycoprotein, α1, transferrin) and fetuin, and its efficiency was independent of the polymeric structure of the molecule or the amount of glycosylation. The method has the additional advantage of using 20% less enzyme than previous methods, which substantially reduces costs.

Preclinical evaluation of light-activatable, bispecific anti-human CD3 antibody conjugates as anti-ovarian cancer therapeutics

The administration of anti-CD3 antibodies, either unmodified or in bispecific formats, has been shown to kill tumors. However, their activity needs to be carefully controlled. We have approached this problem by inhibiting their anti-CD3 activity until it is required. Folated anti-human CD3 antibody bispecific conjugates were therefore synthesised in which the folate portion of the conjugates remained free to bind to folate receptor (FR) expressing cancer cells, whilst their anti-CD3 activity was reversibly inhibited. On irradiation with UV-A light, the T-cell binding activity of the anti-CD3 antibody can be restored only when and where it is required, i.e., adjacent to a tumor. Conjugate bound to FR expressed on normal tissues in other parts of the body remains inactive. This report describes the preclinical in vivo testing of these conjugates in transgenic mice whose T-cells express human CD3 molecules. When the ‘cloaked’ conjugates were reactivated in the region of the primary tumor, both primary tumor growth and liver metastasis were markedly reduced. That the deliberate targeting of T-cell activity locally to the primary tumor also resulted in reduced distant metastatic growth was a key finding. Light-activatable bispecific antibody conjugates similar to those described here offer a means to control T-cell targeting with a much higher degree of specificity to tumors because they minimize potentially dangerous and unwanted side effects in non-illuminated areas. The addition of light-specific targeting to the inherent tumor specific targeting of therapeutic antibody conjugates could result in the development of safer treatments for patients.

The construction of a functional photoactivatable cancer targeting bispecific antibody conjugate.

Given their excellent potential, obtaining therapeutic antibodies which can efficiently target tumour cells has proved to be very difficult. The large majority of antibodies raised against tumour antigens, even those which are quite specific for certain tumours, have some specific or nonspecific cross reactions with normal tissues. This seriously restricts their use in tumour targeting. Photoactivatable antibodies and antibody conjugates could be employed to solve this problem. An antibody which becomes active, only where it is required, on illumination with UV light, would make targeting much more regionally specific. This beneficial effect could be maximised by utilising bispecific antibodies. In these a tumour-specific antibody is coupled to a second antibody capable of binding a substance which is directly or indirectly cytotoxic. If only the cytotoxic end of the antibody were to be reversibly-inactivated, then the tumour-specific end would remain free to bind to its target tumour cells without untoward damage to peripheral normal tissues. Localised illumination of the tumour-targeted bispecific antibody would then maximise tumour destruction whilst minimising damage to other tissues. Herein, we describe the construction of such a bispecific conjugate. An antitumour (carcinoembryonic antigen [CEA] specific) antibody coupled to a photoactivatable monoclonal antibody which binds the enzyme alkaline phosphatase (AP). An anti-AP was chosen as AP has been suggested as a possible enzyme for use in cancer targeting, is easy to detect, and thereby simplifies the analysis of the anti-AP antibody conjugates. The anti-AP antibody was first reversibly deactivated with a 1-(2-nitrophenyl)ethanol (NPE) coating. An average of ten NPE residues coated each anti-AP antibody molecule. When tested in an ELISA the NPE coated anti-AP had approx 30% of the activity of the uncoated antibody, however this increased to approximately 75% of the native value after 10 min irradiation with UV-A light from a hand held 6W UV-lamp. The NPE coupling and uncoupling reactions are well documented and the reaction schemes are given in two previous publications. A major advantage of this procedure is that this low level of light does not damage the antibody or cells (see below). Many previous ‘caging’ procedures require either organic solutions and/or very high levels of irradiation for reactivation. Both of which are very damaging to biological molecules and cells. The NPE coated antibody was then added to AP coated sepharose beads to remove the 30% of the antibody which remained active. After absorption, the ability of the coated antibody to bind AP reduced markedly to as little as 0.5% of its uncoated value (Table 1) with between 50 and 87% of the initial monoclonal antibody activity recoverable after UV irradiation.