Monoclonal antibodies and their tumor-associated antigens
The specific application of monoclonal antibodies to radioguided surgery has been the basis for, and has represented the most important component to, the development of the radioimmunoguided surgery (RIGS) system [
30,
57]. This system was pioneered at The Ohio State University in the early 1980s by the collaboration of a surgical oncologist, Dr. Edward W. Martin, Jr., and a professor emeritus of electrical engineering, Dr. Marlin O. Thurston [
58,
59].
The production of a monoclonal antibody is the result of a technique called hybridoma fusion technology [
60]. Most simply stated, a B-cell lymphocyte (which recognizes a single particular antigen and subsequently produces a single antibody targeting that specific antigen) and a myeloma cell are fused together to create a hybridoma cell. This immortalized hybridoma cell has the ability to survive and replicate outside of the animal. Such a hybridoma cell is able to replicate and be maintained in cell culture and will produce large amounts of a single antibody, which is referred to as a monoclonal antibody.
Monoclonal antibodies used in RIGS can be targeted against antigens expressed on the surface of tumor cells or targeted against antigens expressed within the extracellular environment around tumor cells [
30,
49,
57]. When radiolabeled with various radionuclides, such resulting radiolabeled monoclonal antibody conjugates can potentially be utilized in both diagnostic gamma camera imaging and gamma probe detection of tumors, as well as in cancer therapeutics. In this regard, both whole monoclonal antibodies and monoclonal antibody fragments have been investigated.
The most advantageous features of an ideal monoclonal antibody are: (1) high affinity for its antigen (i.e., the initial ability to bind to the antigen); (2) high avidity for its antigen (i.e., the ability of the antibody to remain bound over an extended period of time); (3) rapid penetration into the tumor tissue; (4) rapid clearance from the bloodstream; (5) minimal accumulation within normal tissues; and (6) the absence of a human antimouse antibody (HAMA) response [
8,
17,
30,
58,
59,
61].
Nevertheless, the production of radiolabeled monoclonal antibody is not necessarily a simple endeavor [
26,
30]. The conjugation of a radionuclide to a monoclonal antibody may potentially change the specific binding properties of the monoclonal antibody. In such an instance in which the specific binding properties of the monoclonal antibody are significantly altered, the resultant radiolabeled monoclonal antibody may be left with significantly reduced affinity and/or avidity for the intended target antigen that ultimately renders the resultant radiolabeled monoclonal antibody clinical ineffectual.
The particular form of the monoclonal antibody (i.e., whether it is a whole monoclonal antibody or a fragment of a monoclonal antibody) can influence its ability to localize tumor [
30]. Monoclonal antibody fragments have smaller molecular weight, have more rapid penetration into tumors, and have more rapid clearance rate from the bloodstream. As a result, the use of radiolabeled monoclonal antibody fragments can result in lower normal tissue background activity and lead to increased tumor to background ratio and improved tumor detections. However, monoclonal antibody fragments tend to accumulate more within the kidneys and, as a result, they may not be useful in the evaluation of the tumors within or around the area of the kidneys or the bladder.
Numerous radiolabeled monoclonal antibodies have been clinically investigated for radioimmunodetection and in RIGS [
30,
49]. The most intensely investigated and clinically evaluated monoclonal antibodies have been those directed against tumor-associated glycoprotein-72 (TAG-72), carcinoembryonic antigen (CEA), and tumor-associated antigen 17-1A. Several generations of anti-TAG-72 monoclonal antibodies have been developed, including two murine-derived anti-TAG-72 monoclonal antibodies (B72.3, native murine CC49) and one humanized anti-TAG-72 monoclonal antibody (HuCC49).
TAG-72 is a tumor-associated glycoprotein with a molecular weight of greater than 10 million Daltons [
62,
63]. TAG-72 contains approximately 80% carbohydrates, has mucin-like biochemical and biophysical properties similar to colonic, small intestine, and gastric mucins, and is thought to be secreted by epithelial tissues [
62,
63]. Numerous epithelial-derived cancers, including colorectal, breast, gastric, pancreatic, ovarian, and non-small cell lung cancers overexpress TAG-72 [
62,
64]. TAG-72 is predominantly located within mucin pools of the extracellular environment around the tumor cells and is not specifically expressed on the tumor cell surface. Of particular importance, TAG-72 has been shown to be associated with over 90% of the colorectal, gastric, and ovarian carcinomas and in approximately 70% of breast carcinomas [
65‐
68]. Finally, while it is rarely expressed in normal human adult tissues or in benign disease processes, TAG-72 is also expressed in some normal human fetal tissues, including normal fetal intestine [
69].
B72.3 was the first-generation murine anti-TAG-72 monoclonal antibody that was developed and was interestingly first derived from reaction with human mammary tumor cells [
70]. B72.3 was shown to be reactive with a variety of human carcinomas, including colorectal (94%), breast (84% of invasive ductal), ovarian (100% of common epithelial), as well as the majority of gastric, pancreatic, endometrial, and lung adenocarcinomas [
30,
65,
67‐
69,
71]. In contrast, B72.3 was shown to have only a very weak or a nonreactivity status to a variety of normal adult human tissues [
30]. The only exception to this rule has been demonstrated for normal postovulatory (secretory phase) endometrium which was shown to be reactive to B72.3, in contrast to normal preovulatory (proliferative phase) endometrium which was nonreactive [
30,
71].
Native murine CC49 was the second-generation murine anti-TAG-72 monoclonal antibody that was developed [
30,
63,
72,
73]. Native murine CC49 was found to have only minimal reactivity to a variety of normal human tissues, recognized a different epitope on the TAG-72 as compared to B72.3, and exhibited higher reactivity than B72.3 to a variety of human carcinomas, including colorectal, breast, ovarian, and lung carcinomas [
30,
72,
73]. From a clinical perspective and as will later be discussed in the clinical application section, native murine CC49 was also superior to B72.3 in tumor detection in RIGS for colorectal carcinoma [
74,
75].
It is well characterized that a majority of patients will develop some degree of a HAMA response to the administration of murine monoclonal antibodies [
30,
64,
76‐
78]. Despite the fact that the HAMA response has been well characterized, its clinical impact on cancer patients, whether deleterious or beneficial, remains very unclear [
79,
80]. Nevertheless, in order to attempt to eliminate this antiimmunoglobulin response, a third-generation humanized anti-TAG-72 monoclonal antibody (HuCC49) was genetically engineered [
81]. HuCC49 demonstrated equivalent tumor-targeting for human colon carcinoma xenografts but a tradeoff of slightly less relative affinity to TAG-72 as compared to native murine CC49 and chimeric CC49 [
81]. However, HuCC49 was shown to not produce a HAMA response [
82]. Further refinements were made in HuCC49 by the development of a higher affinity HuCC49 possessing a CH2 domain deletion (i.e., HuCC49ΔC
H2) [
83]. HuCC49ΔC
H2 demonstrated a more rapid blood clearance, a higher affinity constant (5.1 × 10
-9 versus 2.1 × 10
-9), and significantly lower percent of the injected dose in normal tissues compared to intact HuCC49 [
83], thus indicating the potential utility of the HuCC49ΔC
H2 monoclonal antibody for diagnostic and therapeutic clinical applications. Furthermore, population pharmacokinetic modeling studies have demonstrated that HuCC49ΔC
H2 had more rapid clearance (65% increase) from bloodstream and a resultant shorter "residence time" (24% shorter) than that of native murine CC49 [
84].
Carcinoembryonic antigen (CEA) represents another well-studied and potentially useful target antigen for which radiolabeled monoclonal antibodies have been developed and investigated for RIGS [
30]. CEA is a tumor-associated glycoprotein with a molecular weight of approximately 200,000 Daltons [
85,
86]. It is highly expressed on the cell surface of both embryonic colonic mucosa as well as a wide range of human adenocarcinomas, including colorectal, gastric, pancreatic, breast, ovarian, endometrial, and lung [
30,
85‐
87]. Specific to colorectal adenocarcinomas, it has been previously reported that anywhere from 66% to 100% express CEA [
30].
Numerous murine monoclonal antibodies have been developed to target CEA [
30,
85,
88‐
93]. Those most well studied have included COL-1, A
5B
7, IMMU-4, and CL58. COL-1 monoclonal antibody was first derived from reaction with LS-174T human colon carcinoma xenograft in athymic mice, has a very high affinity to CEA, and has been shown to have a high reactivity to significant number of colon, breast, and lung carcinomas [
30,
85,
88,
89]. Likewise, A
5B
7, IMMU-4, and CL58 represent three additional anti-CEA murine monoclonal antibodies that have shown clinical relevance by possessing a high reactivity to CEA-producing malignancies [
30,
88,
90‐
93].
Lastly, 17-1A (also called EpCAM) is a tumor-associated glycoprotein with a molecular weight in the range of approximately 30, 000 to 40,000 Daltons [
94‐
96] which is thought to represent a cell-cell adhesion molecule. It was first characterized on a human colorectal adenocarcinoma cell line SW1083 [
97]. It is broadly distributed in normal epithelial tissues and in various carcinomas, including colorectal, gastric, and breast [
94,
95,
98].
Murine monoclonal antibodies against the tumor-associated antigen 17-1A were originally developed in the hybridoma SW1083-17-1A [
57,
99,
100]. The localization and clearance properties of the 17-1A murine monoclonal whole antibody and its monoclonal antibody fragment were previously evaluated in a mice xenograft model by Martin et al [
101], demonstrating high tumor-to-normal tissue ratios with highest tumor-to-normal tissue ratios seen at 72 hours and 24 hours, respectively, for the 17-1A murine monoclonal whole antibody and monoclonal antibody fragment [
57,
101].
The most common challenges facing the utility of monoclonal antibodies in radioimmunodetection relate to the activity ratio between tumor and normal surrounding tissues and the time interval between the initial administration of the radiopharmaceutical agent and performance of diagnostic gamma camera imaging or radioguided surgical detection. In an attempt to increase the activity ratio between tumor and normal surrounding tissues and to decrease the time interval between the initial administration of the radiopharmaceutical agent and performance of diagnostic gamma camera imaging or radioguided surgical detection, pretargeting strategies for monoclonal antibodies and radionuclides have been investigated [
102]. Most such pretargeting strategies utilize the principle of the avidin-biotin binding system. This avidin-biotin pretargeting strategy allows for the complete temporal separation of the systemic administration of the monoclonal antibody from that of the systemic administration of the radionuclide. The monoclonal antibody is labeled with biotin and the radionuclide is labeled with avidin. This will ultimately result in a reduction of nonspecific binding. The biotin-labeled monoclonal antibody is first administered, allowing binding of the biotin-labeled monoclonal antibody to the tumor and allowing the nonspecific uptake of the biotin-labeled monoclonal antibody to be cleared. The avidin-labeled radionuclide is then administered and resultantly localizes in the tumor secondary to the high affinity and specificity of the avidin-labeled radionuclide for the biotin-labeled monoclonal antibody. More recently, an additional pretargeting strategy utilizing a bispecific antibody and radiolabeled bivalent hapten system has been investigated that bind cooperatively to target cells [
103].
99mTc-labeled radiopharmaceutical agents
Numerous
99mTc-labeled radiopharmaceutical agents have been formulated for use in diagnostic nuclear medicine by radiolabeling the radionuclide
99mTc to various compounds [
26,
104]. The list of compounds that have been radiolabeled with
99mTc for diagnostic nuclear medicine use is extensive and includes, in alphabetical order, antimony trisulfide colloid, bicisate dihydrochloride, colloidal human albumin (i.e., nanocolloid), colloidal rhenium sulfide, dextran, diethylenetriaminepentaacetic acid (DTPA)-mannosyl-dextran, disofenin, hydroxyl-ethyl starch, exametazime, gluceptate, glucoheptonate, hexakis-2-methoxy-isobutyl-isonitrile (methoxyisobutylisonitrile, MIBI, or sestamibi), hydroxymethylene diphosphonate (HMDP or oxidronate), hydroxyethylene diphosphonate (HDP), lidofenin, mebrofenin, mertiatide (mercaptoacetylglyclyglyclyglycine), methylene diphosphonate (MDP or medronate), pentetate (diethylenetriaminepentaacetic acid), sodium pertechnetate, sodium phytate (D-myo-inositol 1,2,3,4,5,6-hexakisphosphate dodecasodium), sodium pyrophosphate, stannous phytate, succimer, sulfur colloid, teboroxime, tetrofosmin, and tin colloid. The primary
99mTc-labeled radiopharmaceutical agents that have been used for radioguided SLN biopsy include
99mTc sulfur colloid,
99mTc colloidal human albumin, and
99mTc antimony trisulfide colloid. The primary
99mTc-labeled radiopharmaceutical agents that have been used for tumor detection during radioguided surgery include
99mTc MIBI (sestamibi),
99mTc diphosphonates, and
99mTc sodium pertechnetate. The application of
99mTc-labeled monoclonal antibody fragments, such as
99mTc-labeled arcitumomab (IMMU-4 murine monoclonal antibody fragments against CEA) and
99mTc-nofetumomab merpentan (monoclonal antibody fragment of the pancarcinoma murine antibody NR-LU-10) have been used in nuclear medicine imaging but have only been very limitedly investigated for tumor detection during radioguided surgery [
105‐
107].
18F-fluorodeoxyglucose (18F-FDG)
Malignant tumors have long been known to have an accelerated rate of glucose metabolism and have an increased rate of glucose transport and glucose utilization [
110‐
112]. The mechanism of
18F-FDG within malignant cells is well described in the literature [
113‐
115].
18F-FDG is an
18F-labeled nonphysiologic analog of glucose.
18F-FDG within the bloodstream is transported into cells (both malignant cells and normal cells) by a facilitated diffusion mechanism involving specific glucose transporters (i.e., GLUT transporters). Once within the cell,
18F-FDG is phosphorylated to
18F-FDG-6-phosphate by the enzyme hexokinase. Unlike
18F-FDG,
18F-FDG-6-phosphate can not be readily transported across the cellular membrane of either malignant cells or normal cells. The enzyme glucose-6-phosphatase is responsible for dephosphorylating
18F-FDG-6-phosphate to
18F-FDG. Because the enzyme glucose-6-phosphatase is present in relatively low amounts within malignant cells and within normal cells,
18F-FDG-6-phosphate cannot be readily dephosphorylated back to
18F-FDG once
18F-FDG has been phosphorylated within the intracellular environment. Therefore, once
18F-FDG is transported into the malignant cell or the normal cell via the GLUT transporters and is subsequently phosphorylated, the resultant
18F-FDG-6-phosphate is essentially trapped within the cell. Additionally,
18F-FDG-6-phosphate cannot be utilized in the metabolic steps of glycolysis, and, this further lends to the accumulation of
18F-FDG-6-phosphate within the cell. This entire process is thought to occur more readily in malignant cells than in normal cells due to the overexpression of the glucose transporters GLUT 1 and GLUT 3 by malignant cells and due to higher levels of hexokinase within malignant cells. The overall result of this entire process of an accelerated rate of glucose metabolism and an increased rate of glucose transport and glucose utilization by malignant cells is that of a relatively greater accumulation of
18F-FDG-6-phosphate within malignant cells as compared to normal cells. Even more simply stated, malignant cells are much more efficient at accumulating glucose molecules within their intracellular environment than are normal cells. This elegantly elucidated process represents the overall basis for the clinical application of
18F-FDG for the detection of tumor by both diagnostic PET imaging and gamma detection probe technology.
However, limitations do exist in regards to the utilization of
18F-FDG in both diagnostic PET imaging and gamma detection probe technology. These limitations are: (1) the accumulation of
18F-FDG within certain normal tissues with an elevated rate of glucose metabolism (most striking in the brain and heart, and to a lesser degree in the mucosa and smooth muscle of the stomach, small intestine and colon, as well as in thyroid, liver, spleen, and skeletal muscle); (2) the accumulation of
18F-FDG within in inflammatory/granulomatous processes and infectious processes; (3) the excretion and accumulation of
18F-FDG within the urinary tract (kidneys, ureters, and bladder); and (4) the impaired uptake of
18F-FDG in patients with elevated blood glucose levels and with impaired glucose metabolism [
113,
116,
117].