Visualization of immunotoxin-mediated tumor cell death in vivo.

We present a novel methodology to visualize tumor cells directly in a whole mouse. This technique combines immunohistochemistry with whole mouse sectioning. It lets one see the exact distribution of tumor cells throughout an animal and how effectively these cells are eliminated by cancer therapeutics. We used this technique to assess the efficacy of a T cell-specific immunotoxin in a severe combined immunodeficient mouse model of human T-cell leukemia. Severe combined immunodeficient mice were injected with one of two human T-cell acute lymphoblastic leukemia cell lines (Molt 3 and Molt 13) and were either left untreated or were treated with DA7, an immunotoxin specific for the T cell-associated antigen CD7. Mice were sacrificed after tumor cell injection and immunotoxin therapy, whole mouse cross-sections were prepared, and tumor cells in the sections were visualized by immunohistochemistry. No tumor cells were detected in DA7-treated mice injected with Molt 3, consistent with the long-term survival of this group and the sensitivity of Molt 3 to DA7 in vitro. In contrast, DA7 treatment did not visibly eliminate tumor cells in mice challenged with Molt 13, nor did it result in their long-term survival. Furthermore, tumor cells were detected in areas that may have otherwise been overlooked, and their distribution differed from that of mice injected with Molt 13 alone. These analyses indicate that whole mouse sectioning will be a valuable tool for assessing residual disease in the preclinical evaluation of cancer therapeutics.

Identification of a highly conserved module in E proteins required for in vivo helix-loop-helix dimerization.

Basic helix-loop-helix (bHLH) transcription factors often function as heterodimeric complexes consisting of a tissue-specific factor such as SCL/tal or MyoD bound to a broadly expressed E protein. bHLH dimerization therefore appears to represent a key regulatory step in cell lineage determination and oncogenesis. Previous functional and structural studies have indicated that the well defined HLH domain is both necessary and sufficient for dimerization. Most of these studies, however, have employed in vitro systems for analysis of HLH dimerization, and their implications for the requirements for in vivo dimerization remain unclear. Using multiple approaches, we have analyzed bHLH dimerization in intact, living cells and have identified a novel domain in E proteins, domain C, which is required for in vivo dimerization. Domain C, which lies just carboxyl-terminal to helix 2 of the HLH domain, represents the most highly conserved region within E proteins and appears to influence the in vivo conformation of the adjacent HLH domain. These results suggest that HLH dimerization in vivo may represent a complex, regulated process that is distinct from HLH dimerization in vitro.

An in vitro model for toxin-mediated vascular leak syndrome: ricin toxin A chain increases the permeability of human endothelial cell monolayers.

Vascular leak syndrome (VLS) is the dose-limiting toxicity observed in clinical trials of immunotoxins containing ricin toxin A chain (RTA). RTA itself is thought to cause VLS by damaging vascular endothelial cells, but the exact mechanism remains unclear. This is partially due to the paucity of appropriate models. To study VLS, we developed an in vitro model in which human umbilical vein-derived endothelial cells were first grown to confluence on microporous supports and then cultured under low pressure in the presence or absence of RTA. Endothelial cell barrier function was assessed by measuring the volume of fluid that passed through each monolayer per unit time. We found that RTA significantly increased monolayer permeability at times and concentrations consistent with the onset of VLS in patients treated with RTA-based immunotoxins. Scanning electron microscopy showed that intercellular gaps formed in endothelial monolayers exposed to RTA. Intercellular gap formation followed endothelial cell death caused by the enzymatic activity of RTA. We conclude that RTA is directly toxic to endothelial cells in vitro and speculate that this contributes to VLS in vivo.

Construction and characterization of human CD7-specific single-chain Fv immunotoxins.

To develop novel therapeutic agents for treatment of human T cell malignancies, we constructed two single-chain Fv (sFv) immunotoxins specific for the T cell-associated Ag CD7. The sFv fragments were derived from the murine hybridomas 3A1e and 3A1f and were expressed as soluble proteins in Escherichia coli. Surface plasmon resonance analyses demonstrated that the purified 3A1e and 3A1f sFv fragments specifically bound CD7 with high affinity, 8.1 and 1.8 nM, respectively. The difference in affinity is chiefly due to a slower dissociation rate for the 3A1f sFv fragment. Despite this difference, both monovalent sFv fragments were comparably internalized by CD7+ human T leukemic cells within 30 min. These data support findings of previous studies suggesting that CD7 internalization does not require cross-linking. The sFv immunotoxins were assembled by linking ricin toxin A chain to the C termini of the sFv fragments via disulfide bonds. Both sFv immunotoxins were comparably potent in their ability to inhibit protein synthesis in vitro in CD7+ Jurkat cells (50% inhibiting concentration = 15 pM). Further preclinical studies on the use of the 3A1e and 3A1f sFv immunotoxins to treat human T cell diseases therefore appear warranted.

High-level secretion of two antibody single chain Fv fragments by Pichia pastoris.

The diagnostic and therapeutic applications of antibody single-chain Fv (sFv) fragments often require large amounts of protein that can be problematic and expensive to obtain. Here we report the secretion of two sFv fragments by the yeast Pichia pastoris at levels up to 250 mg/l. Soluble sFv fragments were purified from culture supernatants in one step by affinity or metal-chelating chromatography, and were indistinguishable from their bacterially expressed counterparts in terms of affinity. Secretion of functional sFv fragments by Pichia pastoris provides a low cost, high yield alternative to current sFv expression systems.

Platelet factor 4 binds to glycanated forms of thrombomodulin and to protein C. A potential mechanism for enhancing generation of activated protein C.

Platelet factor 4 (PF4) is an abundant platelet alpha-granule heparin-binding protein. We have previously shown that PF4 accelerates up to 25-fold the proteolytic conversion of protein C to activated protein C by the thrombin.thrombomodulin complex by increasing its affinity for protein C 30-fold. This stimulatory effect requires presence of the gamma-carboxyglutamic acid (Gla) domain in protein C and is enhanced by the presence of a chondroitin sulfate glycosaminoglycan (GAG) domain on thrombomodulin. We hypothesized that cationic PF4 binds to both protein C and thrombomodulin through these anionic domains. Qualitative SDS-polyacrylamide gel electrophoresis analysis of avidin extracts of solutions containing biotinylated PF4 and candidate ligands shows that PF4 binds to GAG+ but not GAG- forms of thrombomodulin and native but not Gla-domainless protein C. Quantitative analysis using the surface plasmon resonance-based BIAcoreTM biosensor system confirms the extremely high affinity of PF4 for heparin (KD = 4 nM) and shows that PF4 binds to GAG+ thrombomodulin with a KD of 31 nM and to protein C with a KD of 0.37 microM. In contrast, PF4 had no measurable interaction with GAG- thrombomodulin or Gla-domainless protein C. Western blot analysis of normal human plasma extracted with biotinylated PF4 demonstrates PF4 binding to protein C in a physiologic context. Thus, PF4 binds with relative specificity and high affinity to the GAG- domain of thrombomodulin and the Gla domain of protein C. These interactions may enhance the affinity of the thrombin.thrombomodulin complex for protein C and thereby promote the generation of activated protein C.

Bacterial expression and characterization of an anti-desipramine single-chain antibody fragment.

Tricyclic antidepressant toxicity is a leading cause of death from intentional drug overdose. Monoclonal antibody Fab' fragments specific for the tricyclic antidepressant, desipramine, reverse acute drug toxicity but may themselves have adverse effects at therapeutic doses. To evaluate the characteristics of smaller antibody fragments, we cloned, expressed and characterized a 26 kD single chain Fv fragment (G5-sFv). A DNA sequence encoding VH-linker-V1 was constructed from hybridoma mRNA encoding a high affinity monoclonal desipramine specific IgG1 and expressed in E. coli. G5-sFv was produced at high levels as insoluble inclusion bodies. Single chain Fv was solubilized, folded in a redox buffer and affinity purified on desipramine-Sepharose. The affinity of G5-sFv for desipramine was similar to that of the corresponding monoclonal Fab' as measured by surface plasmon resonance (Fab' 5.5 +/- 0.5 x 10(8) M-1, sFv 2.3 +/- 0.5 x 10(8) M-1). G5-sFv administered to rats after a tracer dose of 3H-desipramine produced rapid and marked redistribution of drug from tissues into serum. G5-sFv was stable at 4 C for greater than 6 months but lost activity at higher temperatures. We conclude that desipramine-specific-single chain Fv expressed in E. coli retains the affinity of the parent antibody for desipramine. The pharmacokinetic effect of G5-sFv on desipramine distribution suggests that it may be useful as an antidote for desipramine overdose.

Laboratory preparation of a deglycosylated ricin toxin A chain containing immunotoxin directed against a CD7 T lineage differentiation antigen for phase I human clinical studies involving T cell malignancies.

An immunotoxin consisting of a monoclonal antibody specific for CD7, a cell surface determinant expressed on T acute lymphocytic leukemia (T-ALL) blast cells, was linked to the potent plant toxin deglycosylated ricin toxin A chain (dgRTA) and is currently under evaluation in phase I clinical trials. Scale-up production of this immunotoxin, called DA7, was simplified using a two-step purification protocol that resulted in a highly purified immunotoxin meeting FDA criteria for IND approval. The anti-CD7 antibody, 3Ale, an IgG2b, was coupled to toxin using two different heterobifunctional cross-linkers, (1) N-succinimidyl-3-(2-pyridyl-dithiolproprionate) (SPDP), considered a standard croslinker and (2) 4-succinimidyloxycarbonyl-alpha-methyl-alpha-(2-pyridyldithio)tolu ene (SMPT), designed to hinder the in vivo breakdown of the toxin/antibody disulfide bond. Since experiments revealed that SPDP-DA7 had similar pharmacokinetics and biodistribution in mice and higher yields than DA7 made with a hindered cross-linker, SPDP-DA7 was scaled up for clinical study. Yield of SPDP-DA7 was 25% relative to starting material. Fractions were collected containing a toxin: antibody ratio of 1:1 to 4:1 rather than only a 1:1 ratio since studies showed that this heterogenous fraction was just as toxic to proliferating CD7-expressing leukemia cells as a homogeneous 1:1 fraction. In vitro, the concentration of heterogenous SPDP-DA7 selectively inhibiting 50% activity (IC50) of the CD7+ CEM cell line was 0.01 microgram/ml to 0.05 microgram/ml for inhibiting activated T cells or T cell lines. In vivo, SPDP-DA7 showed a significant anti-tumor effect against CEM cells administered to scid/scid mice, but even more importantly was effective against primary T cell leukemias taken from patients and injected into scid/scid mice.

Frequent aberrant immunoglobulin gene rearrangements in pro-B cells revealed by a bcl-xL transgene.

During B lymphocyte development, pro-B cells that fail to rearrange an immunoglobulin heavy (IgH) chain allele productively are thought to undergo developmental arrest and death, but because these cells are short-lived in vivo they are not well characterized. Transgenic mice expressing the apoptosis regulatory gene bcl-xL in the B lineage developed large expansions of pro-B cells in bone marrow. V(D)J rearrangements in the expanded populations were nearly all nonproductive, and DJH rearrangements were enriched for joints in DH reading frame 2 and for aberrant joints with extensive DH or JH deletions. Thus, the death of pro-B cells with failed immunoglobulin rearrangements occurs by apoptosis, and bcl-xL can deliver a strong survival signal at the pro-B stage. This analysis also demonstrated that immunoglobulin gene rearrangement is less precise than previously appreciated.

Selection for S107-V11 gene expression by peritoneal B cells in adult mice.

Peritoneal B-1 cells in adult mice express a restricted repertoire of V genes. To determine the basis for the nonrandom expression of one VH gene (the V11 gene in the S107 family), and to determine if this gene is nonrandomly expressed early in development, we amplified rearranged S107 genes in adult and neonatal tissues taken from B10.H-2aH-4bp/Wts (2a4b) mice. In adult peritoneal B cells, the S107-V11 gene was found to be productively rearranged up to five times more frequently than V1, another member of the S107 VH gene family. This did not occur in adult spleen, or in neonatal spleen or liver. In these three tissues the productive S107-V11/V1 rearrangement ratios ranged from 0.5 to 0.9. This difference suggests that S107-V11-expressing peritoneal B cells in adult mice have been selected and clonally expanded. Clonal expansion is further suggested by the differing ratios of productive to nonproductive rearrangements. In adult peritoneal B cells, the S107-V11 productive/non-productive ratio is higher than that of V1, consistent with the clonal expansion of S107-V11-expressing cells but not of V1-expressing cells. Analysis of the VHDJH junctional sequences in the productive S107-V11 rearrangements reveals that there are no predominant restrictions in the lengths or sequences of the third complementarity determining regions (CDR3). Selection is therefore independent of CDR3 and so must be based on germline-encoded portions of the S107-V11 region.

Development and characterization of three recombinant single chain antibody fragments (scFvs) directed against the CD19 antigen.

Antibodies that recognize antigens restricted to leukemia, lymphoma, and normal hematopoietic cells represent a unique opportunity to develop therapeutics, because they have the potential for relatively selective treatment of these diseases. Antibodies that recognize the CD19 antigen found on normal and malignant B cells, but not on stem cells, have been used to develop immunoconjugates. However, these conjugates are large and might be suboptimal in tumor penetration when compared to molecules using smaller single chain Fv (scFv) antibody fragments. scFv has the advantage of being a molecularly engineered homogeneous molecule. In this report, we demonstrate the cloning, expression, and binding of three anti-CD19 antibodies as scFvs. All three scFvs were successfully cloned and expressed. FVS191, derived from cell line B43, and FVS192, derived from SJ25C1, were properly refolded and bound CD19 antigen in FACS competition assays. These anti-CD19 scFv should be useful in the further development of diagnostic and therapeutic molecules.

Development of B-1 cells: segregation of phosphatidyl choline-specific B cells to the B-1 population occurs after immunoglobulin gene expression.

Adult mice have two easily recognizable subsets of B cells: the predominant resting population of the spleen, called B-2, and those called B-1, which predominate in coelomic cavities and can express CD5. Some antibody specificities appear to be unique to the B-1 population. Cells expressing antibody specific for phosphatidyl choline (PtC) are the most frequent, comprising 2-10% of peritoneal B cells in normal mice. To understand the basis for the segregation of the anti-PtC specificity to this population, we have produced transgenic (Tg) mice expressing the rearranged VH12 and V kappa 4 genes of a PtC-specific B-1 cell lymphoma. We find that VH12-Tg and VH12/V kappa 4 double-Tg mice develop very high numbers of PtC-specific peritoneal and splenic B cells. These cells have the characteristics of B-1 cells; most are CD5+, and are all IgMhi, B220lo, and CD23-. In the peritoneum these cells are also CD11b+. In addition, adult mice have many splenic B cells (up to one third of Tg+ cells) that express the VH12 Tg but do not bind PtC, presumably because they express a V kappa gene other than V kappa 4. These cells appear to be B-2 cells; they are CD23+, CD11b-, IgMlo, B220hi, and CD5-. Thus, mice given either the VH12 Tg alone or together with the V kappa 4 Tg develop a large population of PtC-specific B cells which belong exclusively to the B-1 population. Since B-2 cells can express the VH12 and V kappa 4 gene separately, we interpret these data to indicate that the events leading to the segregation of PtC-specific B cells to the B-1 population in normal mice are initiated after Ig gene rearrangement and expression. These data are discussed with regard to hypotheses of the origin of B-1 cells. We also find that VH12-Tg mice have a marked decrease in the generation of Tg-expressing B cells in adult bone marrow, but not newborn liver. We speculate that this may be related to positive selection of VH12-expressing B cells during differentiation.

High frequency expression of S107 VH genes by peritoneal B cells of B10.H-2aH-4bP/WTS mice.

Our previous analyses of peritoneal Ly-1 B cells indicate that a high percentage express VH genes of the VH11 and VH12 families, and that this bias is due to clonal selection. The antibodies encoded by these genes bind the same hapten, phosphatidyl choline (PtC). Twenty-one of 73 hybridomas generated from fusions with peritoneal Ly-1 and Ly-1 sister population B cells of B10.H-2aH-4bp/Wts mice produce anti-PtC specific antibodies. We show here that 19 of these express VH11 and VH12 family genes and two express VH36-60 family genes. To assess whether there is a bias in VH gene use among non-PtC-specific hybridomas we analyzed the remaining 52 hybridomas for VH family expression by using VH family-specific probes in an RNA dot blot assay and by Ig mRNA sequencing. We find a seven-fold increase in the expression of the VHS107 family genes, and only slight differences in the expression of VH genes of other families relative to splenic B cells. We attribute the increase in VHS107 gene expression to clonal selection inasmuch as five of the seven VHS107+ hybridomas express the same VH gene (V11) and VL association is nonrandom. The bias in VH gene use among the entire panel of 73 peritoneal hybridomas is to the extent that approximately one-third express one of three genes: the V11 gene of the S107 family, the CH34 gene of the VH11 family, and the VH12 family gene.

Antibodies specific for Ig idiotype, but not isotype, can substitute for antigen to induce IgM secretion by a B cell clone.

Cells of the mouse B cell clone, CH12.LX, secrete IgM when cultured with nominal antigen (sheep erythrocytes, SRBC) and mAbs which bind their membrane Ek molecules. To determine whether anti-Ig antibodies can substitute for antigen in the induction of IgM secretion by CH12.LX, the B cells were cultured with anti-Ek mAbs and anti-IgM or anti-idiotype antibodies. Anti-IgM antibodies were capable of cross-linking the membrane IgM of CH12.LX, and inhibited mitogen-induced differentiation of the B cells. However, anti-IgM could not substitute for SRBC in delivering a major histocompatibility complex-restricted differentiative signal. In contrast, either polyclonal or monoclonal antibodies specific for the CH12.LX Ig idiotype were fully capable of substituting for antigen in the induction of IgM secretion by CH12.LX. The binding of anti-IgM antibodies did not prevent anti-idiotype antibodies from delivering a differentiative signal. Thus, binding of ligand to different parts of the membrane Ig molecule can result in the delivery of different biological signals to the B cell.

Organization and expression of VH gene families preferentially expressed by Ly-1+ (CD5) B cells.

Previously we have shown that a high proportion (5%-15%) of peritoneal Ly-1+ B cells that bind the common membrane phospholipid, phosphatidylcholine, express either of two VH genes. One of these genes, CH34VH, belongs to the recently described VH11 family, whereas the other VH gene, CH27VH, is reported here to belong to a new, single-member family which we designate VH12. We have characterized these new VH families in terms of strain polymorphism, map position relative to nine other VH gene families, and frequency of expression in polyclonally stimulated splenic B cell populations. We find that VH11 and VH12 genes are expressed at frequencies similar to the VH genes of other families in the spleen, and that these genes are not located among the J-proximal VH segments which are preferentially expressed during early development. These results suggest that the VH11 and VH12 genes are not preferentially rearranged, and support our earlier studies which suggested that the high frequency of VH11 and VH12 expression among peritoneal Ly-1 B cells is due to antigen selection.

Biased immunoglobulin variable region gene expression by Ly-1 B cells due to clonal selection.

Most, if not all, autoantibodies specific for bromelain-treated mouse erythrocytes recognize the common membrane phospholipid, phosphatidyl choline (PtC). Anti-PtC antibodies are produced by 5%-15% of CD5+ Ly-1 B cells of normal unimmunized mice, but not by detectable numbers of conventional CD5- B cells. At 1 week of age PtC-specific B cells are undetectable but then increase dramatically over the next 3 to 4 weeks to reach adult numbers. We report here that PtC-specific Ly-1 B cells in B10.H-2aH-4bp/Wts mice predominantly express either of two heavy and kappa chain variable (V) region gene combinations. In addition, the sequence and length of DH genes are conserved among cells expressing the same V gene combination, and the V kappa-J kappa junctions of one group involve unusual splice sites. Preferential V gene rearrangement models are insufficient to explain the DH and V kappa-J kappa junctional sequences or the delayed appearance of this specificity, and so they cannot solely account for the high frequency of PtC-specific cells. These characteristics are more consistent with antigen selection. We therefore attribute the frequent use of the two V region gene combinations to selection for cells that express them and conclude that the expressed V gene repertoire of Ly-1 B cells in adult mice is influenced by antigen selection. Apparently, there is no selection for mutant anti-PtC antibodies of higher affinity during the formation of the Ly-1 B repertoire because the V region genes expressed by PtC-specific cells are unmutated. Our findings are consistent with an important, germ line-encoded function for the immunoglobulin products of these gene combinations.

Restricted Ig variable region gene expression among Ly-1+ B cell lymphomas.

The majority of the characterized Ly-1+ B cell lymphomas of B10.H-2aH-4bp/Wts origin (the CH series) bear surface Ig related by Ag specificity or idiotype or both. To determine the genetic basis for these structural similarities, we have sequenced the VH and VL region genes expressed by 10 CH lymphomas, and have compared their VH and V kappa gene rearrangements by Southern blot analysis to one another and to those of four other CH lymphomas. Sequence analysis identified only five different VH, and seven different VL genes, and indicated that these V genes are essentially unmutated. CH lymphomas which express the identical VH gene share at least one idiotope. Thus, the basis for shared idiotype and specificity is due in most cases to the use of the same V gene. This restriction in V gene expression is not due to the preferential use of V genes of any particular VH family or VL group, as the expressed V genes belong to four different VH families and four V kappa groups, and include V lambda 1 and V lambda 2. We hypothesize that Ag selection accounts for the restriction in V gene usage among CH lymphomas.