John Shively, Ph.D.

The Shively Lab specializes in studying the CEA gene family. More specifically, he researches the use of anti-CEA antibodies for tumor imaging and therapy, as well as the role of CEACAM1 in T cell activation and mammary epithelial cell polarization. Dr. Shively is also exploring the relationships of the immune system and genetics to fibromyalgia.

Fibromyalgia Research Fund

The City of Hope Fibromyalgia Research Fund supports a specific research project on fibromyalgia that began around 2005 as a research study that enrolled patients from the fibromyalgia clinic of Dr. R. Paul St. Amand AMD collecting blood samples from FMS patients and their parents.

The protocol was approved by the City of Hope Institutional Review Board (IRB) in 2005 [protocol number 04186] and has continued to accrue patients and their parents ever since. The title of the project is “Immunological and genetic analysis of autoinflammatory genes in fibromyalgia.” Collection of blood samples allowed the investigators to analyze (1) the presence of certain proteins called cytokines in the blood that reflect the nature and degree of immune activation of that patient with FMS and (2) collect DNA from the immune cells in the blood and sequence genes that are associated with autoinflammation with the hope that mutations in those genes would correlate with the disease.

Furthermore, the investigators would determine if any mutations discovered were inherited at a higher frequency than expected, thus requiring the analysis of their parents’ genes. To facilitate enrollment in this project, the investigators no longer require blood samples, but instead collect saliva for genetic analysis only.

The Carcinoembryonic Antigen Gene Family

The CEA gene family comprises 30 genes located on chromosome 19. There are two subgroups or clusters, one includes CEA, NCA, BGP, CGM1, CGM2 and CGM6, and the other includes the PSGs (pregnancy specific glycoproteins). The first group are cell surface glycoproteins linked to the membrane either as type I transmembrane proteins (e.g., BGP, CGM1) or via a GPI (glycosylphosphatidylinositol) moiety (e.g., CEA, NCA, CGM2, CGM6). The second group are glycoproteins secreted by the placenta or fetus, hence the term "pregnancy specific." CEA was named for its discovery in fetal colon and colon carcinomas, but apparent absence in normal adult colon. NCA (nonspecific crossreacting antigen) and BGP (biliary glycoprotein) were found in normal tissues, especially in neutrophils and epithelial cells of the digestive tract. Later studies identified further members of the family designated "CGMs" for CEA-gene-man for the most part also expressed in neutrophils and epithelial cells (CGM6 is restricted to neutrophils). All are members of the Ig superfamily and have similar domain structures. While many in vitro studies have shown that CEA gene family members can function as homophilic cell adhesion molecules, their in vivo functions are poorly understood. In fact, in light of their apical expression in epithelial cells, we strongly doubt that they play a role in homotypic cell adhesion. Another possible function includes bacterial binding as exemplified by the recent finding that the N-domain of most of the family members bind Nisseria meningitis and gonorrhea. While the lack of an in vivo function has hampered studies in this field, the use of radiolabeled anti-CEA antibodies to target tumors of the colon, breast, ovary and lung has become an important tool for tumor imaging and therapy.

Functional Studies on BG

We have performed functional studies on BGP in three cell systems, the induced expressed of BGP in activated human T cells, the constitutive expression of BGP in the normal human mammary cell line MCF10F under morphogenic conditions, and the transgene expression of BGP in the murine colon carcinoma cell line MC38. Activation of T cells with anti-CD3 or PHA (phytohemagglutinin) plus IL-2 induces the expression of BGP (CD66a) over a 2-5 da period. Expression levels of CD66a are inversely correlated to CD25 (IL2R?) expression and fall when IL2 is removed from the media. CD66a is expressed as two isoforms, both with the long cytoplasmic tail that has the ITIM (immunoglobulin tyrosine phosphate inhibitory motif). When the cells are stimulated with anti-CD66a antibody, the ITIM is phosphorylated, suggesting that CD66a crosslinking delivers an inhibitory growth signal. Immunoprecipitation of CD66a co-precipitates TCR (T cell receptor) associated molecules ZAP-70, Lck, and Vav. In addition, Shp-1, known to associate with the ITIM of BGP in epithelial cells, is co-precipitated, as are actin, myosin and calmodulin. Current studies are aimed at linking these signaling pathways to functional studies.

MC-38 murine colorectal cells transfected with human BGP shown a strong association of BGP with cytoskeleton (actin-myosin filaments) after activation with anti-BGP antibodies and phosphorylation of the ITIM on BGP. Two dimensional gel electrophoresis analysis of the BGP immunoprecipitates from anti-BGP treated cells reveal 10-12 spots that have been analyzed by in situ trypsin digestion and LC/MS/MS (liquid chromatography/mass spectrometry/mass spectrometry) analysis. Current studies are using GST-BGPcyt fusion proteins to determine if there is a direct association between the BGP cytoplasmic domain and the cytoskeleton, or if linking proteins are involved.

The MCF10F mammary epithelial cell line is capable of forming acini when grown in serum free conditions either in or on matrigel, a basement membrane mimic. While these cells express BGP before and after exposure to matrigel, the BGP expression is apical (luminal) in mature acini. If the cells are sorted into BGP high or low populations and grown on matrigel, the BGP high cells form acini exclusively, while the BGP low cells form mixtures of tubules and acini. We have also shown that MCF10F cells can yield a population of BGP negative, spindle shaped cells that have the phenotype of myoepithelial cells. The myoepithelial cells form web-like structures when grown alone, or surround the epithelial cells when grown in mixtures, resembling structures found in the mammary gland. These studies were originally prompted by the observation that BGP is down regulated in over 90% of colorectal cancers, but in only 30% of breast cancers. It has been postulated that since BGP is a product of fully differentiated cells and delivers a negative growth signal, its expression cannot be tolerated in tumor cells. This concept appears to be only partially true in breast cancer, and in the studies shown here, we suggest that BGP expression may occur early during morphogenesis without disrupting acini and tubule formation. Current studies are aimed at knocking out the expression of BGP in these cells to determine its possible role in differentiation.

Anti-CEA antibodies for tumor imaging and therapy

CEA is an excellent target antigen for cancers of the colon, breast, and lung. We have developed the anti-CEA antibody T84.66 which has a high affinity (1010 M-1) and specificity for CEA suitable for in vivo tumor targeting. We have conjugated the engineered mouse-human chimeric version of this antibody and its fragments with a variety of chelates including DTPA (diethyltriamine pentaacetic acid) and DOTA (tetraazacyclododecyltetraacetic acid) for radiometal labeling. Current efforts have focused on improving the biodistribution properties of the radiolabeled chelate-conjugates by increasing the metal conjugate stability and by introducing linkers that are chemically labile, allowing for greater blood clearance after tumor localization. The macrocycle DOTA has proven ideal for these studies. It has a high binding constant (10 23 M-1) for 111In(III) and 90Y(III), radiometal ions that have excellent emissions and half lives suitable for imaging and therapy, respectively. We have coupled DOTA to the amino acid cysteine, followed by addition of the homobifunctional, sulhydryl specific crosslinker BMH (bis-maleimidohexane) to form MC-DOTA which allows site specific conjugation to the cysteines in the hinge region of antibodies. When radiolabeled and injected into tumor bearing animals, the conjugate targets tumors with high uptake (>50% injected dose/g tumor) and has good blood clearance due to the slow breakdown of the succinimide bonds in the conjugate. We are now focusing on improving the rates of radiometal binding using both chemical and theoretical approaches. The latter approach involves a collaboration with Dr. William Goddard’s group at Caltech have performed ab initio calculations on DOTA-Y(III) complexes. We are now planning to scan theoretical combinatorial structures using this approach, followed by verification of the best structures through chemical synthesis and NMR studies.