Peptide priming of cytolytic activity to HER-2 epitope 369-377 in healthy individuals.

The presence of tumor-reactive CTLs in tumor infiltrates and in the peripheral blood of cancer patients demonstrates an immune response against tumors that apparently cannot control disease spread. This raises concerns as to whether amplification of this response may be useful during disease progression. Induction of tumor-reactive CTLs in healthy donors at risk, as well as in patients free of disease, may be therapeutically important, based on the hypothesis that CTLs that recognize tumors early may be more effective in containing their progression than CTLs that expand only when the disease progresses. To address the feasibility of priming cytolytic activity in healthy donors, we used the HER-2 peptide E75 (369-377) as an immunogen and autologous peripheral blood mononuclear cell-derived dendritic cells as antigen-presenting cells. We found that of 10 healthy donors tested, two responded at priming with E75 presented on autologous dendritic cells by induction of E75-specific CTL activity. Three other responders were identified after two additional restimulations. Of these five responders, three recognized E75 presented on the ovarian tumor line SKOV3.A2, as demonstrated by cold-target inhibition experiments. Induction of cytolytic activity at priming was enhanced in responders by tumor necrosis factor-alpha and interleukin 12 but not in the nonresponders. AlphaB7.1 monoclonal antibody added at priming enhanced induction of lytic activity in only one of the four nonresponding donors tested, suggesting that in the majority of donors, E75-precursor CTLs were not tolerized. Because of the possibility that disease may develop in nonresponders, strategies to improve the immunogenicity of tumor antigens for healthy donors may be required for development of cancer vaccines.

Induction of determinant spreading and of Th1 responses by in vitro stimulation with HER-2 peptides.

Immunization with tumor antigens induces cellular and humoral immune responses. These responses by T cells are specific for defined epitopes (determinants) in the molecule of the immunizing tumor antigen. Extension of such responses to self-antigens requires induction of autoimmunity to the tumor. As with systems of autoimmune disease, expression of T cell autoimmunity is charaterized by diversification of responses from the inducer determinant to other responder (cryptic) determinants. Since similar strategies may be useful for therapy of human cancers, we investigated whether the induction of response to a HER-2 peptide F7 (776-789) induces enhanced reactivity of other HER-2 peptides. We found that stimulation with F7 can expand a response to another epitope F13 (884-899) in both an ovarian cancer patient with progressive disease and a healthy donor who shared HLA-DR11. This response was characterized mainly by increased interferon gamma secretion, and proliferation, but was not observed with another donor who shared HLA-DR14 and HLA-DQ5 with the patient. Since repeated vaccination with the same epitope may lead to a decline of primary cell reactivity caused by apoptosis spreading the response to other epitopes, the tumor antigen may provide an approach for maintaining an inflammatory Th1 response during cancer vaccination.

Identification of activated tumor antigen-reactive CD8+ cells in healthy individuals.

We investigated the ability of HER-2 peptide E75, which maps an immunodominant CTL epitope for ovarian and breast tumor-associated lymphocytes (TAL), to activate effector functions in freshly isolated CD8+ cells from healthy individuals. IFN-gamma was rapidly induced by E75 within 20-24 h, in five of six healthy donors, in the presence of IL-12 and was detectable as early as 6 h. The IFN-gamma levels were Ag-concentration dependent. Similar results were obtained with peptides mapping CTL epitopes from two other tumor Ag: folate binding protein (FBP) and amino-enhancer of split of Notch (AES). IFN-gamma was also detected, from freshly isolated, unstimulated PBMC in response to HLA-A2 matched tumors + IL-12 but not of IL-12 alone. The major source of IFN-gamma were CD45RO+ CD8+ cells. Induction of IFN-gamma and IL-2 from CD8+ cells and of IL-12 from dendritic cells (DC) by CD8+ cells reactive with E75 mirrored their induction by the influenza matrix peptide (M1: 58-66) in the same individual. Responses to M1 are used to define the presence of activated memory cells in healthy individuals. Compared to M1 responses E75 recognition induced 2-4-fold lower levels of IL-12 from the same APC and IFN-gamma and IL-2 from the same CD8+ cells. At lower Ag concentrations the endogenous IL-12 induced by E75-reactive CD8+ cells did not reach the threshold required to co-stimulate for IFN-gamma. alphaB7.1 synergized with E75 in increasing the overall levels of IL-2 induced within 24 h. The presence of tumor Ag-reactive activated CD8+ cells in healthy individuals may improve our understanding of the mechanisms of immunosurveillance and regulation of immune responses by tumors.

Vaccine implications of folate binding protein, a novel cytotoxic T lymphocyte-recognized antigen system in epithelial cancers.

The immune system can be efficiently stimulated and targeted to specific antigens expressed exclusively or preferentially by experimental cancers. The foremost limitations to extending this vaccine technology to the prevalent epithelial-derived cancers are the lack of: (a) identified tumor-associated antigens recognized by cellular immunity; (b) antigens expressed on the majority of tumor cells during disease progression; and (c) immunogenic CTL epitopes. To date, only HER-2/neu has been shown to be the source of naturally occurring, MHC-restricted, CTL-recognized peptides in epithelial tumors. In this study, we demonstrate that the human high-affinity folate binding protein (FBP), which is a source of antigenic peptides recognized in ovarian cancer, is also recognized in breast cancer. Both immunodominant E39 (FBP, 191-199) and subdominant E41 (FBP, 245-253) epitopes are presented by HLA-A2 in these cancers. These peptides are efficient at amplifying the response of tumor-associated lymphocyte populations in terms of lytic function, enhanced proliferation, and specific IFN-gamma release. On a per cell basis, tumor-associated lymphocytes stimulated with the FBP peptides exhibit enhanced cytotoxicity not only against peptide-loaded targets but also against FBP-expressing epithelial tumors of different histologies. Furthermore, FBP peptides induced E39-specific CTLs and E39- and E41-specific IFN-gamma and IP-10 secretion in certain healthy donors. The broad distribution of FBP among >90% of ovarian and endometrial carcinomas, as well as 20-50% of breast, lung, colorectal, and renal cell carcinomas, along with pronounced differential overexpression in malignant tissues compared with the extremely limited expression in normal epithelium, suggests the exciting potential of a widely applicable FBP-based vaccine in epithelial cancers.

Folate binding protein peptide 191-199 presented on dendritic cells can stimulate CTL from ovarian and breast cancer patients.

Tumor associated lymphocytes (TAL) isolated from malignant ascites cultured in media containing interleukin-2 show antitumor responses. These antitumor responses are mediated by cytotoxic T lymphocytes (CTL) which recognize antigen in the context of MHC molecules using T cell receptors. CD8+ CTL recognize peptide epitopes processed from cellular proteins in the context of MHC class I molecules. These peptides have a restricted length of 8-11 amino acids. The folate binding protein (FBP) is overexpressed in over 90% of ovarian and 20-50% of breast cancers. We recently found that FBP is the source of antigenic peptides recognized by a number of these CTL-TAL. This indicated that FBP peptides are antigenic in vivo for ovarian and breast CTL-TAL. To define FBP immunogenicity, a peptide defining the epitope E39 (FBP, 191-199) was presented by PMBC derived dendritic cells (DC) from healthy donors isolated by the CD14 method to ovarian and breast CTL-TAL. Stimulation of ovarian and breast CTL-TAL by E39 pulsed DC (DC-E39), in the presence of IL-2, rapidly enhanced or induced E39 specific CTL activity. This E39-responder population consisted of cells expressing TCR V beta 9, V beta 13, and V beta 17 families, based on the increase in the percentages of these families in DC-E39 versus DC-NP stimulated TAL. Characterization of immunogenic tumor antigens and of cytokine requirements for induction of functional antitumor effectors may be important for future cancer vaccine developments.

Ovarian cancer-associated lymphocyte recognition of folate binding protein peptides.

BACKGROUND: Tumor-associated lymphocytes (TAL) isolated from ovarian cancer patients contain cytotoxic T lymphocytes (CTL) capable of recognizing specific HLA/peptide complexes on tumor cells leading to tumor cell lysis. Currently, HER2/neu, overexpressed in only 30% of breast and ovarian cancers, is the only known source of CTL-recognized peptides in epithelial cancers. Therefore, we have investigated peptides derived from folate binding protein (FBP), which is over-expressed in more than 90% of ovarian cancers and in the majority of other epithelial tumors. METHODS: TAL were isolated from the malignant ascites of four consecutive HLA-A2+ ovarian cancer patients and incubated in IL-2. Initial chromium-release assays were performed within 1 week. T2 cells, incubated with peptide, were used to reconstitute T cell epitopes. The FBP sequence was interrogated for HLA-A2 binding peptides, and five were synthesized (E37-41). RESULTS: Freshly cultured, unstimulated ovarian TAL recognize peptides derived from FBP. These peptides are presented in the context of HLA-A2, and are specifically recognized in a HLA class I-restricted fashion. TAL recognition of these reconstituted T cell epitopes is concentration dependent. Furthermore, the FBP peptides are shown by cold target inhibition studies to be naturally processed and presented antigens. CONCLUSIONS: FBP peptides are recognized by freshly isolated TAL from ovarian cancer patients, suggesting in vivo expression and sensitization. Because FBP is over-expressed 20-fold in most adenocarcinomas, these peptides may be used in a widely applicable peptide-based vaccine for epithelial tumors.

Proliferative and cytokine responses to class II HER-2/neu-associated peptides in breast cancer patients.

Previous studies have characterized the reactivity of CD8+ CTLs with ovarian and breast cancer. There is little information about the antigens and epitopes recognized by CD4+ T cells in these patients. In this study, we analyzed the ability of T cells from peripheral blood mononuclear cells of breast cancer patients to recognize HER-2/neu (HER-2) peptides. We found that 13 of 18 patients responded by proliferation to at least one of the HER-2 peptides tested. Of these peptides, one designated G89 (HER-2: 777-789) was recognized by T cells from 10 patients. Seven of nine responding patients were HLA-DR4+, suggesting that this peptide is recognized preferentially in association with HLA-DR4. Analysis of the specificity and restriction of the cytokine responses to G89 by G89-stimulated T cells revealed that these cells secreted significantly higher levels of IFN-gamma than interleukin 4 and interleukin 10, suggesting priming for a Th0-T helper 1 response. The same pattern of cytokine responses was observed to the intracellular domain of HER-2 protein, suggesting that G89-stimulated T cells recognized epitopes of the HER-2 protein in association with HLA-DR4. Because HLA-DR4 is present in 25% of humans, characterization of MHC class II-restricted epitopes inducing Th0-T helper 1 responses may provide a basis for the development of multivalent HER-2-based vaccines against breast and ovarian cancer.

Wilms' tumor with renal cell carcinoma overgrowth in a 3-year-old child.

Wilms' tumors that contain features of both renal cell carcinoma and classic Wilms' tumor histology are rare. Even though nine such cases have been previously reported in the literature, we report the first case of a Wilms' tumor with an overwhelmingly renal cell carcinoma histologic pattern.

Ovarian and breast cytotoxic T lymphocytes can recognize peptides from the amino enhancer of split protein of the Notch complex.

In this study we investigated recognition by ovarian tumor associated lymphocyte (OVTAL), and breast tumor associated lymphocytes (BRTAL), of peptides corresponding to the sequence 125-135 of the Aminoenhancer of split (AES) protein. Three of these peptides designated as G75:AES1/2 (128-135), G60: AES1/2 (127-137) and G61: AES1/2 (125-133) correspond to the wildtype AES sequence, while the fourth G76:GPLTPLPV, AES1/2 (128-135) corresponds to a variant sequence of the peptide G75 with the N-terminal Leu substituted to glycine. These sequences were chosen for study because mass-spectrometric analysis (MS) of a CTL active HPLC peptide fraction eluted from immunoaffinity precipitated HLA-A2 molecule, revealed: (a) the presence of an ion with a mass-to-charge ratio (m/z) of 793 which was more abundant than other ions of similar masses; (b) the tentatively reconstituted sequence of the ion 793 matched the sequence of peptide G76. We found that AES peptides G75 (128-135) and G76 (128-135) (L128G) reconstituted CTL recognition at concentrations ranging between 200-500 nM. These concentrations are lower than concentrations reported to activate effector function of CTL recognizing other epithelial tumor Ag. Furthermore, analysis with cloned CD8+ T cells indicated that G75 and G76 were not cross-reactive specificities, suggesting a key role for the N-terminal residues of the variant peptide in dictating specificities. Since the AES proteins are part of a set of transcriptional repressors encoded by the Enhancer of split [E(spl)] genes, and since these repressors are activated to suppress cell differentiation in response to Notch receptors signalling, the AES peptides may represent a novel class of self-antigens that deserve further consideration as tumor Ag in epithelial cancers.

Identification of naturally processed human ovarian peptides recognized by tumor-associated CD8+ cytotoxic T lymphocytes.

Identification of naturally processed peptides recognized by tumorspecific CTLs may lead to epitope-specific tumor vaccines. Because these epitopes may be expressed differently on epithelial tumors and may differ in their ability to induce CTL in vivo, we have isolated the HLA-A2-peptide complexes by immunoaffinity from an established ovarian tumor line transfected with and expressing HLA-A2 gene. High-performance liquid chromatography-fractionated peptides were used to reconstitute epitopes recognized on HLA-A2 by three HLA-A2+ CD8+ CTL lines. These lines recognized at least three of the same groups of fractions (designated SKOV3.A, -B, and -C) but showed differences in the pattern of recognition of other fractions. To gain insight in the epitope distribution by freshly isolated ovarian tumors, we compared the recognition of peaks SKOV3.B and -C with the corresponding peaks from an ovarian tumor (OVA-6) that expressed similar levels of HLA-A2, using one of these lines (CTL-OVA-5) as indicator. CTL-OVA-5 recognized a large number of epitopes from peaks B and C rechromatographed on more resolving high-performance liquid chromatography gradient. Although a number of peaks appeared to be coincident on both SKOV3 and OVA-6, an even higher number appeared either not to overlap or to overlap only partially. These findings, which represent the first analysis of the epitopes presented by a patient tumor, suggest that the use of tumor line-derived peptides for vaccination may require selection of the epitopes corresponding to the ones presented by freshly isolated human tumors.

Physiological measurements of luminal stirring in the dog and human small bowel.

The resistance to absorption resulting from poor stirring of luminal contents (RLum) is considered to be equivalent to an unstirred layer of greater than 600 microns in the human small intestine. We measured RLum in the jejunum of conscious dogs by assessing the absorption rate of two rapidly absorbed probes, glucose, and [14C]warfarin. When RLum was expressed as an unstirred layer, the maximal thickness of the unstirred layer (assuming negligible epithelial cell resistance) was only approximately 35 and 50 microns for perfusion rates of 26 and 5 ml/min, respectively. Maximal unstirred layer thickness for the human jejunum, calculated from previous studies of glucose absorption, yielded a mean value of only 40 microns (range: 23 to 65 microns). Since epithelial resistance appears to be negligible during absorption of low concentrations of glucose, the maximal unstirred layer of 40 microns should be close to the true value for glucose in the human small intestine. We conclude that the unstirred layer for rapidly absorbed compounds in dogs and man are less than one-tenth of previously reported values, but this layer still may remain the rate limiting step in absorption of rapidly transported compounds.

Influence of infusate viscosity on intestinal absorption in the rat. An explanation of previous discrepant results.

Previous studies of the influence of increased luminal viscosity on intestinal absorption have yielded conflicting results ranging from no effect to a marked diminution. We measured the absorption of three probes (carbon monoxide, [14C]warfarin, 5.5 mM glucose) from a saline infusate or from saline containing 0.6% guar, which yielded a 20-fold increase in viscosity. Two animal models were used: (a) conscious nonlaparotomized rats with chronically implanted cannulas and (b) anesthetized laparotomized rats. In the anesthetized laparotomized rats, absorption was independent of perfusate viscosity. In the conscious nonlaparotomized rats, the absorption of each of the three probes was significantly greater than in the anesthetized laparotomized rats and increased viscosity caused a 60%-70% decrease in the clearance of the three probes. In anesthetized laparotomized rats, we have shown that fluid moves with laminar flow, and increased infusate viscosity cannot further reduce luminal stirring (or absorption). In conscious, nonlaparotomized rats, laminar flow is disrupted by normal gut motility causing better luminal stirring. Such stirring is inhibited by a viscous infusate resulting in decreased absorption. We conclude that the conflicting results seen in previous studies can be attributed to the model used. In conscious animals where luminal stirring was good, a viscous infusate caused decreased absorption.

Role of specific interactions between protein kinase C and triphenylethylenes in inhibition of the enzyme.

The tumor promoter receptor protein kinase C (PKC) has been implicated as a key enzyme in cellular growth regulation. It is, therefore, believed that specific PKC inhibitors may include effective antiproliferative agents. Previously, we have shown that the antiestrogen tamoxifen and related triphenylethylenes are potent inhibitors of PKC. Although the mechanism of inhibition of PKC by triphenylethylenes clearly involves nonspecific interactions between the antiestrogens and the lipid cofactor of PKC, we recently demonstrated that PKC itself has specific triphenylethylene-binding sites, suggesting that the inhibitory mechanism also involves specific drug-protein interactions. In this report, we characterize the direct interactions between PKC and triphenylethylenes and demonstrate their relevance to the inhibitory action of triphenylethylenes against PKC. We show (a) that the triphenylethylene-binding sites of PKC are located in the catalytic domain of the enzyme, (b) that MgATP (i.e., 10 mM MgCl2 plus 1 mM ATP) competes with the triphenylethylenes for binding sites on PKC, and (c) that triphenylethylenes are competitive inhibitors of PKC with respect to MgATP. Taken together, these data provide strong evidence that triphenylethylenes can inhibit PKC by binding directly to the ATP-binding region of the active site of the enzyme. The specific interactions between triphenylethylenes and PKC characterized here may provide a rationale for developing more specific PKC inhibitors.

Physiological measurement of luminal stirring in perfused rat jejunum.

Poor stirring of intestinal contents yields a preepithelial diffusion barrier that is thought to be the rate-limiting step in absorption of many compounds. In many previous studies, the resistance of this barrier is equal to an unstirred water layer of 300-800 micron. Using three probes, CO, glucose, and [14C]warfarin, we measured the preepithelial resistance in a 30-cm segment in rats that were 1) conscious, 2) anesthetized with pentobarbital sodium, or 3) anesthetized and laparotomized. Measurements with each of the probes showed that the maximal preepithelial resistance in conscious rats was equivalent to an unstirred layer of only approximately 100 micron. Anesthesia roughly doubled this resistance, and anesthesia and laparotomy caused a sixfold increase (unstirred layer of approximately 600 micron). We conclude that the luminal stirring is much more efficient than previously has been appreciated. The very thick jejunal unstirred layers reported previously (300-800 micron) reflect the results of studies performed under nonphysiological conditions or studies employing inappropriate techniques to measure luminal stirring.

Suppression of the auditory frequency following response during visual attention.

Frequency-following responses (FFRs) were recorded from unanesthetized cats with electrodes chronically implanted in the cochlear nucleus and on the round window. Tone bursts of different frequencies (irrelevant stimuli) were presented repetitively (85 dB SPL, 1/sec) as background before, during, and after the presentation of a visual discrimination task (relevant stimuli) which attempted to alter the attentive state of the animals. The mean peak-to-peak amplitudes of the FFRs from the cochlear nucleus were significantly reduced in amplitude during attention to the visual discrimination stimuli when compared with the amplitudes of the pretest- and posttest-control periods. However, at the round window (cochlear microphonic) no significant differences in amplitude were observed for the same periods. Although the amplitudes of the FFRs were reduced in amplitude at all frequencies during visual attention, much greater suppression occurred at the middle frequencies (700-2000 Hz) than at higher or lower frequencies. These data suggest that during visual attention the FFRs are attenuated by a central inhibitory mechanism.

Auditory and visual evoked potentials to irrelevant stimuli during conditioning to a visual stimulus.

Amplitude changes of evoked responses to irrelevant auditory (click) and visual (optic tract stimulation) stimuli were examined during appetitive conditioning to a flashing light. Four female cats, with chronically implanted electrodes along the visual and auditory pathways, were conditioned with pairings of a flashing light and food reinforcement. The results showed that, as conditioning progressed, the auditory evoked potentials were attenuated while the visual evoked potentials did not change significantly. The data seem to indicate possible differences in the neural processing of irrelevant information.

Relevant stimuli and auditory evoked potentials.

Four cats, classically conditioned to a flashing light paired with food reinforcement, were tested for amplitude changes of click-evoked potentials during increasing hours of deprivation. Signal averaged evoked potentials from the auditory cortex, cochlear nucleus, and round window electrodes showed no significant changes in amplitude, but significant changes were found in the variance as food deprivation increased. Such changes are more consistent with expected changes in the CNS than the amplitude changes of the auditory cortex found by Saunders and Chabora.