NO169947B - PROCEDURE FOR THE PREPARATION OF A MONOCLONAL ANTIBODY TO A TUMOR ASSOCIATED ANTIGEN - Google Patents

Description

The present invention relates to a method for producing a monoclonal antibody

to a tumor-associated antigen to which a metal radionuclide is bound, and the distinctive feature of the method according to the invention is that the monoclonal antibody is labeled with a metal radionuclide by means of a chelating agent that is conjugated to the antibody, using it as a chelating agent a polyamino carboxylic acid chelating agent of formula

wherein X is 0 or an integer, y is 1 or 2, z is an integer from 1 to 7 and R is hydrogen or a group with which the chelating agent is conjugated to the antibody or a group that affects the binding constant of the chelating agent for the radionuclide.

These features of the invention appear in the patent claim.

A reliable technique for site-specific detection of tumors has long been sought. It is e.g. described methods for the detection of carcino-embryonic antigen (CEA) associated with human carcinoma and which circulates in the blood stream, see US Patent No. 3,663,684 and 3,697,638. These methods clearly cannot be used to determine the tumor location. Recently, it has been proposed to use antibodies labeled with radioactive isotopes of iodine to detect an antigenic substance associated with a tumor. Thus, US patent no. 3,927,193 describes a method in which

anti-substances are turned to CEA marked with and ^1.^

According to the aforementioned patent document, it remained

experiments carried out with male Syrian-type hamsters,

in which human signet-ring cell carcinoma was introduced,

by injection of labeled goat anti-CEA followed by examination of the animal's organs, demonstrated that the radioisotope localized in the tumor. It was therefore proposed that the location of a tumor in a human could be determined by in vivo administration of a parenteral solution of the antibody followed by photoscanning of the host.

The current use of radioiodinated antibodies has not satisfied the expectations that were set for the previous works. E.g. reported Goldenberg et al, N. Eng. J. Med. 298, 1384-1388 (1978) had some success in 1978 when screening people with radioiodinated anti CEA. Because of. however, residual background activity was the reported limited success depending on the use of the subtraction technique. Mach et al, N. Eng. J. Med., 303, 5-10, (1980) reported even less success and determined that 0.1% of the delivered dose localized in the tumor.

In selected cases, however, detection of the tumor was still possible.

It is well known that affinity purification of heterologous antisera to obtain the antibodies used in previous tumor detection processes usually results in the loss of high-affinity antibody and that the antibodies obtained,

which is a mixture of specific antibodies for different determinants on the antigen, mostly comprising antibodies with low affinity and antibodies that give non-specific reactions. Monoclonal antibodies, on the other hand, can be selected to exhibit high affinity to selected sites on the antigen and low non-specific binding. It is

however, surprisingly found that monoclonal antibodies to the turaor antigen labeled with radioisotopes of iodine according to well-known methods still show poor localization at the tumor site despite the fact that in vitro immunoreactivity can be demonstrated. This can be, or probably is, the loss of radioactive iodine from the labeled antibody.

Consequently, there is still a need for a reliable technique for in vivo detection of the exact location of a tumor.

In the present invention, a monoclonal antibody or antibody fragment to a tumor antigen is bound with a chelated radionuclide, e.g. DTPA-bound ]-11In, by conjugation of a chelating agent with the antibody to form a complex between the chelating agent and the radionuclide. The resulting labeled anti-substance, or fragment thereof, when injected into a tumor-bearing host, will quickly localize in the tumor itself with high specificity. In some cases, localization in distant metastases can also occur and indicates spread of the tumor. Localization of the tumor can be detected using photo-scanning techniques. Mixtures of labeled monoclonal antibodies can also be used.

It is therefore an aim of the present invention to obtain monoclonal antibodies to tumor-associated antigens labeled with metallic radionuclides.

The invention enables the detection of tumors in an infected host.

Achievement of these and related purposes will be apparent from the subsequent description of the invention, in connection with the attached drawings, in which: Fig. 1 is a group of graphical representations illustrating the distribution in tissue of<11>1In and 125j compared to

6 7

the distribution of Ga as a function of time after injection in hairless mice infected with hurrian colon cancer

111 125 67

of anti-CEA labeled with ' In and I and Ga citrate.

Fig. 2 is a group of graphic representations that illustrate

111 125 67 increases the amount of In and I and Ga secreted by hairless mice with increasing time after injection of "<*>"^In and "''^I

6 7

labeled anti-CEA and Ga-citrate.

Fig. 3 is a group of graphs showing the tissue distribution of "^^in and "*"^I in hairless mice 48 hours after injection of <11>"1"In and 125I labeled anti-CEA compared to ^" "^In citrate and "^"^1-labeled human albumin. Fig. 4 is a group of graphs illustrating tumor/tissue ratios for <111>ln and 125i compared to <67>Ga, after injection of <11>1In and 125I labeled anti-CEA and ^Ga-citrate in hairless mice, as a function of time.

In accordance with the present invention, monoclonal antibodies or fragments thereof to turn-associated antigens are labeled with chelated radionuclides". The "monoclonal antibodies used in the invention are products of the so-called hybridoma technology which is now well known. It is displayed, e.g. to the original work of Milstein and Kohler, reported in Nature, 256, 495-497

(1975). In principle, the process involves injection

of mice or other suitable animal with an immunogen, in this case a tumor-associated antigen such as e.g.

CEA. The immunized mouse is then killed and cells harvested

from the spleen are combined with myeloma cells to give hybrid cells called "hybridomas" that can be reproduced in vitro. The population of cells that produce antibodies is cultured selectively, and selected to isolate individual

clones, each of which secretes a single antibody species specific to a particular region ("determinant") on the surface of the antigen molecule.

The individual clones can be further selected to determine those that produce antibodies with the highest affinity for the antigen, and that exhibit low non-specific binding. The antibody selected for radiolabeling (usually isolated from ascites) is brought to conjugate with the specified chelating agent which is then used to bind the radionuclide. The chelating agent can be brought to the antibody using a wide variety of methods.

The chelating agents used include a number of polydentate chelating agents where ligands are complexed with metallic radionuclides, in the form of the polyamino-carboxylic acids with the formula:

where X is 0 or an integer (x=0 or x=l is preferred), y=l or 2 (y=l is preferred) and z is an integer from 1 to 7 (z=l or 7 is preferred) and in which R is hydrogen or a group through which the chelating agent is conjugated with the antibody or a group that affects the binding constant of the chelating agent for the radionuclide. Among usable groups R, the group in which A is the group with which the conjugation is achieved can be mentioned in particular. E.g. A can be ~^ 2+ which is obtained by diazotization of the group -NH2, which itself can be obtained by reduction of a nitro group (-NCO which in turn can be obtained by nitration of the phenyl nucleus. The group A can also be

achieved

by acetylation of the group -NH2 using known methods, in which L is a leaving group which is replaced during the conjugation, e.g. a halogen such as Cl, Br or I (Br is preferred).

In other cases, R can be -CH3 or -CH2C6H4OCH2CO2H.

For these polyaminocarboxylic acids, the conjugation can

is achieved via the carboxylic acid group itself by forming e.g. an acid anhydride intermediate taught by Krejcarek and Tucker, Biochemical and Biophysical Research Communications, 77, 581 (1977).

Of the polyaminocarboxylic acids, polyaminoacetic acids are preferred. Specific polyamino carboxylic acids that can be used include ethylenediaminotetraacetic acid (EDTA) and derivatives thereof such as e.g. di-ethylenediamino-pentaacetic acid (DTPA), p-bromoacetamidophenyl (ethylenedinitrilo-tetraacetic acid) and 1-(p-aminophenyl)-ethylenediaminotetraacetic acid) the latter of which is converted to a diazonium salt to allow conjugation. Of these chelating agents, DTPA is preferred.

Radionuclides that can be used in the invention are those that form stable complexes with polydentate chelating agents.

The chelating agents of formula 1 form stable complexes with radionuclides in a +3 oxidation state

state. Among radionuclides that form stable complexes

in +3 oxidation state is<52>Fe,<11:L>In, 113mln<99m>TC<68>Ga, 67Ga, 169Yb, 57Co, 167Tm, 166Tm, 146Gd, 157Dy, 9^b 103Ru, 97Ru, 99Rh, 101mRhog<2>01T1.

The most suitable radionuclides are gamma emitters with an energy between 100 and 400 keV with half-lives in the range from approximately 5 hours to 5 days.

For the production of the monoclonal antibody, the more flexible route of first conjugating a chelating agent to antibodies and then forming a complex with the radionuclide is used. In this way, it is particularly preferred to use<111>In (T1/2=2.8 days) as the radionuclide in conjunction with DTPA as a chelating agent.

The invention can be utilized for the detection of tumors using monoclonal antibodies to tumor

associated antigens in general. In addition to CEA,

can e.g. melanoma-associated antigens, alpha-feto-

protein, ferritin, human chorionic gonadotropin, zinc glycinate marker and prostatic acid phosphatase (PAP) e.g. are used as immunogens to stimulate the production of the desired monoclonal antibodies as described above.

In current use, the labeled antibody in an appropriate parenteral solution is injected into the client. After sufficient time, the client is photoscanned to detect any areas of localized radiation that could indicate a tumor and/or one of its metastases.

The following experiments carried out with monoclonal antibody to CEA labeled with ^2^iQg In demonstrate the advantage of using monoclonal antibodies labeled with chelated radionuclides for tumor detection in accordance with the invention compared to processes using iodine-labeled antibodies.

1. Preparation of radiolabeled anti-CEA.

Monoclonal: anti-CEA, available from Hybritec, Inc. La Jolla, Ca., USA, was labeled with 12 5i using the method of Bolton and Hunter, Biochem. J. 133, 529-39

(1973). The product was passed through a "Sephadex G-25) column and the final material contained less than 0.5% free iodine. Radiolabeled monoclonal antibody binding was more than 80% with horse anti-mouse antibody to sepharose and more than 75% with CEA bound to sepharose.

The same monoclonal anti-CEA was labeled with "'"^In using conjugated DTPA as the chelating agent according to the method of Krejcarek and Tucker, Biochemical and Biophysical Research Communication, 77, 581, (1977). The reaction product was purified by passing through a "Sephadex G-75" column. Antigen binding was comparable to that occurring with 12 5I labeled antibody. The method of labeling did not affect the affinity constant

Ka, for the antibody relative to unchelated antibody

as determined by titration with CEA.

In vitro stabilities of the labeled anti-CEA antibodies were determined by incubating a portion of the material in a parenteral solution comprising sterile normal saline, U.S.P., at 37°C for 72 hours and chromatographic determination of free radionuclide. The antibody solutions were deposited on strips of Baker alumina IB and developed in 50% aqueous acetone. The protein-bound radio-

nuclide remains in place while free radionuclide migrates. After the incubation period, the radioiodinated material showed less than 3% activity present as free iodide. Storage at 4°C reduced this by 50%. Under the same conditions, ^"'"In anti-CEA showed comparable stability.

2. Distribution studies in hairless mice.

Tissue distribution studies of the radiolabeled monoclonal mouse anti-CEA were performed in hairless mice bearing 0.1 to 1.0 gm human colon tumors. The mice weighed from 17 to 28 gm (average about 23 gm). The tumors were initiated by subcutaneous injection of a mixture containing CEA-producing human colon adenocarcinoma cells. Distribution studies were carried out 3 weeks later. Two trials were conducted. In the first, 19 animals were killed

divided into 4 groups of 4 to 6 animals and were injected intravenously (IV) with 0.1 ml of a sterile normal saline

125

solution U.S.P. containing 0.5 microcurie I of anti-CEA antibody (approximately 0.3 micrograms of antibody) and as a control carrier-free 6 7Ga-citrate, a known non-

specific human colon tumor detection agent. Injections were made without anesthesia - using a 100 microliter Hamilton syringe to ensure accuracy. Animals in which the dose partially infiltrated the tail were omitted

from the survey. After injection, the mice were placed in sterile metabolic cages with raised wire mesh floors for collection of urine and feces. Under the floor, sterile 4x4 wire mesh was filled in such a way that it could not be gnawed by the animals. The animals, kept in a clean room at 4°C, with sterile feed and water ad libitum, were euthanized in groups 4, 24, 48 and 72 hours after the injection of the tracer.

The mice were anesthetized with ether and blood samples were obtained by direct cardiac puncture. The mice were then euthanized by cervical dis-localization. During subsequent dissection, the tumour, liver, kidneys, muscle, heart, lung, femur and spleen were removed, - cleaned 2 times in water, 1 time in 10% formalin, dried and weighed wet on an analytical scale. Intact stomach and viscera were also removed and measured. Blood, muscle and bone were assumed to represent respectively

7, 40 and 10% of the total body weight for calculations about

total organ uptake. The other organs were weighed together. Radioactivity determinations were carried out in an auto-gamma particle counter with apertures arranged over the 27-35 keV photopeak of I and the 93 keV photopeak of Ga. The counter was programmed to reject after 10 minutes or count to 1 million to achieve statistical significance for the samples with the lowest activity levels. Corrections for background and reciprocal contribution of isotopic activity between opening settings were made by using prepared sources and solving simultaneous equations. The activity in the samples was then compared with standards for the injected material, with results expressed as % injected dose per grams and % injected dose per organ. Tumor to tissue ratios were calculated from the % dose/gm data. Activity in intestines, urine and faeces was calculated as % of total injected dose.

In the second experiment, 37 mice bearing tumors became hairless

grown from the same human colon cancer cells divided into 5 groups of 7 to 8 animals. One group served as a control and

125

was simultaneously injected IV with one microcurie of I human serum albumin (0.01 mg USA) and 3 microcuries of carrier-free ^^In citrate (0.01 mg citrate) and euthanized 48 hours after injection. In the remaining 4 groups, the animals received simultaneous injections of 3 microcuries of ^^"In-labeled anti-CEA (1.5 micrograms of antibody) and 5 microcuries of carrier-free 6 7Ga-citrate. The groups were euthanized after 4, 24,

48 and 72 hours. Tissue dissection, sampling for radio assessment, counting technique and calculations were carried out as described above. In the case of<111>I-n, the spectrometer was set to count the 247 keV photopeak using a keV aperture (210-250 keV)

but Ga was counted with a 30 keV aperture overlapping the 93 keV photopeak. These settings resulted in less than 10% reciprocal contribution of radioactivity.

As shown in fig. 1, increased the concentrations of ^^^In

anti-CEA which localized in the tumor continuously during the course of the experiment. 23% of the injected dose had accumulated after 72 hours and the results suggest that this level would have been even higher if later sampling had been attempted. The blood level, on the other hand, decreased during the time period. The liver showed a rapid uptake after 4 hours followed by a period of little change and then an increase to 72 hours. The other tissues showed little variation over the investigated time period.

6 7

The content of Ga in the tumor remained fairly constant during the course of the experiment, being about 1/4 of that of "L"L1In anti-CEA after 72 hours. The blood level of 6 7Ga steadily fell while the liver concentration showed a pattern somewhat similar to that "'""'"'''Indium antibody f. Muscle-6 7

the concentration of Ga fell until 48 h after which it leveled off, while the bone concentration remained the same throughout.

The amount of iodine in the tumor peaked after 4 hours

it was initially somewhat greater than the amounts of "'""''"''Indium and then decreased unevenly during the remainder of the experiment, being only 1/10 of"'"''"■''In at 72 hours.At this time, twice as much<67>Ga appeared in the tumor

125 125

as I. Count rates for I in all tissues declined rapidly from their maximum at 4 h and in some cases background levels were reached by 48 h.

The excretion data (fig. 2) show a steady increase in urinary ''""'""''In between 24 and 72 hours and reaches a maximum of 14% of the injected dose. The faecal concentration of<III>In remained fairly constant at 2 4 and 48 h and then increased to 17% at 72 h, while the amount in the viscera remained relatively constant throughout the period

and never exceeded 5% of the dose.

6 7

The excretion pattern of Ga was quite similar to that of

■^■^In men • still showed more trace material in the bowels and faeces than observed on ^ In and less in the urine.

125

Almost 60% of the I radioactivity was removed in

the urine during the first 24 hours and 75% after 48 hours.

125

A further 5% of I was lost in the faeces. After

At 72 hours approximately 80% of <125>I initially injected had been excreted by the animal. Chromatography of 125

In the urine, showed that most of it was free

iodide with another species presumably representative

125

free I complexed to something other than a protein.

Fig. 3 presents data for ^"^In labeled anti CEA at 48 hours and compares it with ^Indium citrate and -*-2^I HSA. ''"In-eit rat was used, as a control for any indium that could have dissociated from antibody,

125

while I HSA was used as a non-specific protein control. In anti-CEA demonstrated a completely different distribution pattern from the "^^In citrate and concentrated more efficiently in the tumor by a factor of 10. Uptake of the two tracers by the other tissues was also different. The tumor concentrations of<125>I H§A was 20% of<11>''"In anti-125 125

CEA, but still 50% greater than for anti-CEA. ii-.HSA distribution in non-tumour tissue is also much different from the corresponding for the ^"^In anti-CEA product. It is particularly

125

significantly that I HSA, which has long been considered a stable iodinated protein, showed almost 60% excretion

(43% in the urine) after 48 hours, remarkably consistent with the previous finding that 12 5I anti-CEA was also

found in the urine. The urinary fraction probably represents free iodine.

The tumor/tissue ratio (Fig. 4) shows ^"^"'"in anti-CEA

as superior to 6 7Ga with respect to all tissues with the exception of blood, which was approx

6 7 12 P

the same as for Ga in this case. In anti-CEA conditions, while they are apparently preferred over them

111T 67- ..

shown by In and Ga in some tissues are misleading as they are written from low<125>I levels in normal tissue rather than from localization in the tumor.

In light of the previous experiments, 12 5I anti-CEA is clear

not suitable as a tumor localizing agent. The rapid and extensive de-halogenation lowers the tumor tracer concentration and increases the blood background and this is prohibitive for its use in detection without sophisticated subtraction techniques. At 72 hours, the tumor thus contains twice the amount of the non-specific<67>Ga at one per gram. base such as<125>I, and furthermore the tumor/blood ratios are better for 6 7Ga than for the iodinated antibody. When one takes into account that 6 7Ga is considered not suitable for imaging adenocarcinoro in the colon, the difficulties that Mach et al,

support on, N.Eng. J. Med. 303, 5-10 (1980). The data

125

which applies I demonstrates unequivocally that in absence

131

of subtraction technique, if I were applied, would

131

very large doses of the labeled material had to be added to achieve the necessary tumor concentrations for effective imaging. This would result in the client receiving a high radiation dose due to it

8-day physical half-life and 608 keV beta-la 4- ' 131X

the component of I.

In contrast, it allows a large percentage of

the injected dose of ^"^in anti-CEA introduced into the tumor and the more favorable physical properties of ^"^in injection of smaller amounts of the radio-pharmaceutical substance, while the imaging is improved and the radiation dose

is reduced for the client.

The tumor/background ratios formed by ^"^In anti-CEA are well within the range required for imaging purposes due to the high degree of radionuclide localization in the tumor. Furthermore, the data suggest that the tumor acquires radiolabeling that would otherwise have been incorporated into or was attached to other tissues.

Individual mice were photoexamined 4, 24, 48 and 72 hours after injection of<111>ln-labeled anti-CEA using a "Phogam H P Anger" gamma camera. After 48 hours, the tumor was well defined without the need to resort to the subtraction technique. After 72 hours the tumor definition was further improved.

It is well known that CEA and some other tumor-associated antigens are secreted or secreted by the tumor. In such cases, the occurrence of background radiation that occurs due to the antigen circulating in the blood or other tissues and combining with labeled antibody will be reduced by initially adding unlabeled antibody to combine with circulating antigen. This will also tend to reduce the higher liver uptake of the labeled antibody. After an appropriate period, the labeled antibody is added and the client is photoscanned to determine possible sites for localized irradiation.

While the preceding discussion has pointed to the use of a single monoclonal antibody labeled with a radionuclide,

it is also possible to use two or even more labeled monoclonal antibodies for different antigenic determinants. The different antibodies are selected to be non-interfering with the binding of the antigen in another antibody. By using several, non-interfering labeled antibodies, tumor imaging can be improved as the antigen will have a higher capacity for

useful for imaging a tumor where associated antigen is present in a low concentration.

Claims (1)

Method for producing a monoclonal antibody to a tumor-associated antigen to which a metal radionuclide is bound, characterized in that the monoclonal antibody is labeled with a metal radionuclide by means of a chelating agent which is conjugated to the antibody, the chelating agent being a polyamino -carboxylic acid chelating agent with formula wherein X is 0 or an integer, y is 1 or 2, z is an integer from 1 to 7 and R is hydrogen or a group with which the chelating agent is conjugated to the antibody or a group that affects the binding constant of the chelating agent for the radionuclide.