Contrasting cellular and humoral autoimmunity associated with latent autoimmune diabetes in adults

OBJECTIVE

Diabetes is clinically classified into two types: type 1 (T1D) and type 2 diabetes (T2D). Nevertheless, intermediate forms of diabetes are frequent and difficult to recognize and manage appropriately. In this study, we investigated whether patients with intermediate form of diabetes, here called unclassified diabetes (UD), have beta-cell autoimmune markers.

RESEARCH DESIGN AND METHODS

beta-cell autoimmune markers (beta-cell autoantibodies (aAb), peripheral blood mononuclear cells (PBMC) responsive to five islet proteins, cytokine secretion, and human leukocyte antigen (HLA)-DQB1 genotypes) were analyzed in 50 UD patients, 23 age- and HLA-matched normal control subjects, and 23 classic T2D patients.

RESULTS

We observed that 16 out of 50 (32%) UD patients demonstrated responsive PBMCs, as opposed to 1 out of 23 (5%) age- and HLA-matched normal control subjects, and 0 out of 23 classic T2D patients. Overall, 29 (58%) UD patients had at least one marker of beta-cell autoimmunity (beta-cell aAb and/or PBMC autoreactivity), in association with high-risk HLA genotypes DQB1*0201 and/or DQB1*0302. Moreover, the 13 (26%) UD patients who had beta-cell aAb were not the same as those with PBMC autoreactivity, except for one patient. Patients with PBMC autoreactivity were older at the onset of the disease and had a better residual beta-cell function than those with beta-cell aAb.

CONCLUSIONS

Our data confirm that T-cell autoimmunity can be detected in latent autoimmune diabetes in adults patients. We show an inverse correlation between humoral and cellular beta-cell autoimmunities. Possible protective cellular responses in the patients with beta-cell PBMC autoreactivity could have potential therapeutic implications.

Human gammadelta T cells express a higher TCR/CD3 complex density than alphabeta T cells

The aim of our study was to compare CD3 expression on gammadelta T cells and alphabeta T cells in human patients. The antigen density of TCR and CD3 on both subsets was assessed by a quantitative method in eight patients. In parallel, we developed and validated a reliable direct tricolor staining protocol that we tested on samples from hospitalized and healthy individuals (n = 60). Our results demonstrate that human gammadelta T cells constitutively express approximately twofold more of the TCR/CD3 complex than alphabeta T cells. We suggest that this enhanced expression of the TCR/CD3 complex could contribute to the higher reactivity of gammadelta T cells compared to alphabeta T cells. These clinical laboratory results confirm the fundamental data described elsewhere. gammadelta T cells deserve further clinical investigations to understand their precise role in human immunity.

[Standardization and automation of HLA B27 typing by flow cytometry: validation and comparison with microlymphocytotoxicity]

One of the strongest known association between human leukocyte antigen (HLA) phenotype and disease is that of ankylosing spondylitis and HLA-B27. Thus, the determination of HLA-B27 status is an useful tool in the diagnosis of ankylosing spondylitis. To date, the 2 reference methods for HLA typing (microlymphocytotoxicity and molecular biology techniques), are costly in terms of both technician time and materials, and require a great deal of experience. In total, these techniques are not well-suited for routine application in clinical immunology laboratories. Use of flow cytometry has recently been applied for HLA-B27 typing. Nevertheless, it requires an extensive validation protocol. We developed a flow cytometry technique as standardized as possible (whole blood, automated lysing system, automated photomultiplier voltage calibration, definition of thresholds stable with time) and validated our results by comparison with microlymphocytotoxicity. In total, 326 samples were analyzed. We found 99% of concordant results between the 2 techniques, and neither false positive results nor false negative results with flow cytometry could be observed. These results illustrate the reliability of the protocol. It should be remembered that reference technique remains necessary to confirm the few results (< 1%) found in "grey zone" by flow cytometry. Standardization of flow cytometry techniques, as described in this work for HLA B27, seems to be a reasonable goal for the next decade in clinical immunology laboratories.

Factors that predict chemotherapy-induced myelosuppression in lymphoma patients: role of the tumor necrosis factor ligand-receptor system

PURPOSE

To analyze factors that predict the occurrence of chemotherapy-induced myelosuppression and, in particular, the role of the tumor necrosis factor (TNF) ligand-receptor system in lymphoma patients at the beginning of their treatment.

PATIENTS AND METHODS

We investigated the predictive factors for myelosuppression after the first course of chemotherapy in a cohort of 101 consecutive, previously untreated lymphoma patients receiving regimens that include doxorubicin and cyclophosphamide. Plasma samples were tested at baseline by enzyme-linked immunosorbent assay for TNF and its soluble receptors. Univariate and multivariate analyses were performed with a forward regression procedure that included all of the parameters that were found to be significant in the univariate analysis. The dose of chemotherapy and the prophylactic treatment with granulocyte colony-stimulating factor were deliberately included in this model.

RESULTS

Sixty-seven patients experienced World Health Organization (WHO) grade 4 neutropenia, and 37 patients experienced febrile neutropenia, which was responsible for WHO grade 2 through 4 infections in 23 patients. In multiparametric regression analysis, the occurrence of grade 4 neutropenia was associated with high doses of cyclophosphamide (odds ratio ¿OR, 19.8; P =.008) and high levels of soluble p75-R-TNF (OR, 8.52; P =.001). The duration of grade 4 neutropenia for more than 5 days was associated with the lack of hematopoietic growth factor administration (OR, 6.76; P =.004) and high levels of soluble p75-R-TNF (OR, 5.84; P =.0023). The occurrence of febrile neutropenia was associated with high doses of cyclophosphamide (OR, 4.7; P =.007), altered performance status (OR, 18.8; P <.0001) and high levels of soluble p75-R-TNF (OR, 3.49; P =.029).

CONCLUSION

This study indicates that in addition to the dose of chemotherapy and the administration of hematopoietic growth factors, poor performance status and high p75-R-TNF levels can predict the occurrence of chemotherapy-induced myelosuppression in lymphoma patients. This model may help in selecting patients for prophylactic growth factor administration.

[Biological diagnosis of immediate drug allergy. Comparative study of histamine liberation tests and CD63 expression by flow cytometry (preliminary results)]

The diagnosis of drug allergy is mainly based upon a detailed clinical history, positive skin tests (ST) and detection of specific IgE. In case of discrepant results, in vitro investigations to identify the responsible drug are needed. The histamine release test (HR) is usually performed. Nevertheless, its clinical benefit remains controversial. Flow cytometric methods (FCM) for the study of allergen induced basophil activation have been recently described. We assessed their usefulness in the diagnosis of drug allergy in comparison to HR results. Eighteen patients were included and 24 drugs (mainly antibiotics and muscle relaxant drugs) were tested. On the basis of clinical signs, 15 patients were classified as allergic (18 drugs). Sensitivity of biological investigations were found as follows: 71% (ST), 71% (FCM) and 24% (HR). This suggests performing FCM rather than HR. In addition, HR is more costly in terms of both reagents and laboratory technician time. Thus, CD63 detection by FCM seems to be a more reliable method in the clinical immunology laboratory. Additional data are needed to validate these preliminary results (sensitivity and specificity, especially in atopic patients) and to assess the interest of the method to investigate drug allergy due to other types of molecules.

Detection of allergen-induced basophil activation by expression of CD63 antigen using a tricolour flow cytometric method

In the field of allergy diagnosis, most in vitro functional tests are focused on basophils. Nevertheless, the very small number of circulating basophils limits these experiments and their clinical benefit remains controversial. As flow cytometry is a valuable tool for identifying cell populations, even at low concentrations, we developed a tricolour flow cytometric method for the study of allergen-induced basophil activation. Identification of cells was based both on CD45 expression and on the presence of IgE on the cell surface, since basophils express high-affinity receptors for IgE (Fc epsilon RI). Cell activation upon allergen challenge was assessed by the expression of CD63 antigen on the plasma membrane. Basophil isolation and activation (with the chemotactic peptide formyl-methionyl-leucyl-phenylalanine) were validated in 32 non-allergic patients. In 12 allergic patients, basophil stimulation by a relevant allergen was in most cases positive (10/12). Furthermore a concentration-dependent hook effect was observed. Of the allergic and non-allergic patients, none showed non-specific activation with an irrelevant allergen (specificity 100%). Overall, our preliminary results, even in a small population, suggest that this is a reliable and valuable method for the diagnosis of allergies complementing specific allergen IgE and skin test results. Obviously, additional clinical studies are needed to validate these first results.

Cytokine assays in human sera and tissues

The use of accurate and sensitive methods for the measurement of cytokines in body fluids is an absolute prerequisite for the proper use of these mediators in clinical practice. Many factors contribute to the complexity of cytokines quantitation: these molecules circulate at very low levels (e.g. pg/ml) under various molecular forms, the existence of circadian rhythms has been described, and the presence of inhibitors (binding proteins, soluble receptors, autoantibodies) can potentially interfere in the assays. Blood collection for cytokines needs particular attention to prevent possible contamination by endotoxins, which can trigger cytokines cellular production after sampling. Bioassays historically preceded immunoassays; the latter techniques are now very popular, but there is an urgent need for standardisation between the different kits commercially available. Nevertheless, due to the essentially local effects of cytokines, the study of their circulating levels only represents the 'tip of the iceberg' and is of limited value for a global understanding of the pathophysiology of these mediators. This explains the development of other approaches to assess the ability of cells to produce cytokines. These include the ELISPOT assay, the measurement of cell-associated cytokines by flow cytometry, and the study of cytokine secretion by isolated peripheral blood mononuclear cells or by whole blood test. All these techniques, associated with a local detection of cytokines by immunohistochemistry or in situ hybridization and reverse transcriptase polymerase chain reaction (RT-PCR), appear to be complementary tools for a better understanding of the multiple aspects of the biology of cytokines.

Elevated circulating levels of TNFalpha and its p55 soluble receptor are associated with an adverse prognosis in lymphoma patients

In 88 newly diagnosed lymphoma patients, tumour necrosis factor alpha (TNFalpha) and soluble TNF type I receptor (p55-R-TNF) were prospectively determined in plasma by immunoradiometric assay (IRMA) and ELISA methods respectively. These 88 patients included 19 with centrocyto-centroblastic lymphoma, 13 patients with other low-grade lymphoma, and 56 with high-grade lymphoma. Median TNFalpha plasma values were 20 pg/ml (range 5-380 pg/ml) in patients versus 7 pg/ml (range 4-9 pg/ml) in 20 healthy control subjects. Presence of TNFalpha level > or = 20 pg/ml was significantly associated with elevated LDH level (P<0.0001), serum beta2-microglobulin level > or = 3 mg/l (P<0.0001), haemoglobin < or = 12 g/dl (P=0.0001), Ann Arbor stage III or IV disease (P<0.005), and with bulky tumour (P=0.01). High level of TNFalpha was also associated with B symptoms (P<0.005), poor performance status (P<0.05), and serum albumin < or = 35 g/l (P<0.05). Levels of p55-R-TNF were also markedly elevated in these lymphoma patients (median of 3.5 ng/ml, range 0.8-18.8 ng/ml) versus 1.45 ng/ml in control subjects (range 1.1-2.3 ng/ml). Level of p55-R-TNF > or = 3.5 ng/ml was significantly associated with poor performance status (P<0.0001), B symptoms (P<0.0001), beta2-microglobulin levels > or = 3 mg/l (P<0.0001), serum albumin < or = 35 g/l (P=0.0001), C-reactive protein > 6 mg/l (P=0.0003), elevated (>20 pg/ml) IL-6 level (P<0.005), haemoglobin < or = 12 g/dl (P<0.005), and bulky tumour (P<0.001). In the whole group of 88 patients, both high TNFalpha and p55-R-TNF levels strongly predicted short progression-free survival (P<0.005 for both variables) and overall survival (P<0.001 and P<0.001 respectively). In multivariate analyses the elevation of p55-R-TNF retained a higher significance over the other variables and therefore improved the predictive value of the International Prognostic Index. This study suggests that elevated TNF gamma and p55-R-TNF levels have high correlation with other adverse prognostic factors in lymphoma patients and may predict a poor outcome.

Anti-tumor activity of CC49-doxorubicin immunoconjugates

The TAG72 reactive monoclonal antibody CC49 was conjugated to doxorubicin with a malonate linker. The immunoconjugate, designated CC49-BAMME-CH-DOX, was approximately a log less potent than unconjugated doxorubicin in an in vitro cytotoxicity assay. Immunoreactivity of the antibody was fully retained. When evaluated in a nude mouse xenograft model with the antigen positive LS174T human colorectal tumor target, CC49-BAMME-CH-DOX and free doxorubicin had similar tumor suppressive activities. The immunoconjugate was clearly less toxic, however, as measured by weight loss and deaths. When evaluated in an NIH:OVCAR-3 human ovarian carcinoma xenograft model, CC49-BAMME-CH-DOX was superior at prolonging survival in comparison to free doxorubicin, unmodified CC49, and a non tumor binding doxorubicin immunoconjugate. These results indicate that targeting of doxorubicin with the CC49 antibody can improve the toxicity and/or the potency of the drug, depending on the tumor target being evaluated. CC49-Doxorubicin immunoconjugates should be considered for clinical evaluation.

[Methods for studying cytokines in biological media]

Accurate and sensitive methods for the measurement of cytokines in biological fluids are an absolute prerequisite for a proper use of these mediators in clinical practice. The authors describe the numerous molecular forms under which cytokines can possibly circulate. The potential influence of inhibitors (autoantibodies, soluble receptors) on cytokine assays is discussed. Blood collection for cytokines needs a particular attention to prevent a possible contamination by endotoxins which can trigger cytokines cellular production after sampling. If bioassays historically preceded immunoassays, these last techniques are now very popular, but there is an urgent need for standardisation between the different kits commercially available. In addition, in situ detection of cytokines using molecular biology is an helpful complementary technique to the measurement of circulating cytokines.

Production of proinflammatory cytokines and cytokines involved in the TH1/TH2 balance is modulated by pentoxifylline

The modulation of cytokine release induced by pentoxifylline (PTX) has recently been demonstrated not to be restricted solely to tumor necrosis factor (TNF)-alpha. This prompted us to study the influence of PTX on a larger spectrum of cytokines with proinflammatory actions [TNF-alpha, interleukin-6, (IL)-6, IL-1 beta, IL-8] or with implied actions in the TH1 (IL-2, IFN-gamma)/TH2 (IL-10) balance. The IL-1RA was also explored. This work was performed using a whole-blood model in which cytokine production is measured after stimulation by lipopolysaccharide (LPS) (25 micrograms/ml) and phytohemagglutinin (PHA) (5 micrograms/ml) in 1:10 diluted whole blood. The stimulation test was performed in blood from healthy controls and from septic patients (without septic shock) in the presence or absence of PTX at 10(-6), 10(-5), 10(-4), or 10(-3) M. In controls and septic patients, at a 10(-4) M PTX concentration the production of IL-2 is strongly diminished (26-32% of the basal level), followed by diminution of IFN-gamma (30-40%). As expected, of the proinflammatory cytokines TNF was the most strongly suppressed (50% of baseline) followed by IL-1 (about 80% of basal production). Finally, IL-10 was also influenced by PTX (65% of baseline). At 10(-4) M, IL-1RA and IL-6 were unaffected by PTX. Taken altogether, our data demonstrate that PTX possesses a much broader spectrum of activity on cytokine production than was initially described, and it appears to be a potential and promising immunotherapeutic agent.

Reduction of toxicity of a vinca alkaloid by an anti-vinca alkaloid antibody

Inadvertent oncolytic overdoses occur rarely, but can have serious consequences. We have investigated the possibility of using an antibody, 27.8.1A, reactive with vinca alkaloids, as a means of reducing the toxicity associated with overdose situations. In vitro cytotoxicity of a vinca derivative, 4-desacetyl- vinblastine-3-carbox-hydrazide (DAVLBHYD), with and without the addition of 27.8.1A, was determined. Using CCRF-CEM, a human acute lymphoblastic leukemia cell line, as a target in this assay, we observed a greater than 90% increase in cell viability using 100 micrograms/ml 27.8.1A with a 0.1 microgram/ml concentration of DAVLBHYD. 27.8.1A had no effect on cell viability when doxorubicin was used as a control drug in this assay. Similarly, the addition of an irrelevant antibody. EGFrL11, had no effect on the toxicity of DAVLBHYD. In an in vivo survival experiment, nude mice were injected with a toxic dose of DAVLBHYD and subsequently given four doses of 27.8.1A. All anti vinca antibody-treated mice survived, in contrast to the untreated group or irrelevant antibody-treated group in which only 25% and 10% of the mice survived, respectively.

Analytical performances of commercial ELISA-kits for IL-2, IL-6 and TNF-alpha. A WHO study

The determination of cytokine levels in biological fluids by accurate and sensitive methods is an absolute need to be able to define the involvement of these mediators in various clinical situations. To test if this goal was achieved with ELISA kits, a comparative study was undertaken using various commercial kits (6 for IL-2, 8 for IL-6 and 9 for TNF-alpha). The major aims of the work were to critically analyze the analytical performances of the kits and to illustrate some of the pitfalls the users may face. Substantial differences were noted in terms of sensitivity and behaviour of commercial standards versus international reference preparations. These results clearly illustrate the urgent need for a real standardization of immunoassays for cytokine quantitation.

Cytokine production during intraperitoneal chemohyperthermia

The aim of the present work is to investigate in vivo cytokine production during chemohyperthermia. 11 patients suffering from gastric adenocarcinoma (n = 6), ovarian adenocarcinoma (n = 4) or malignant mesothelioma (n = 1) were studied. Patients received 60 mg mitomycin or 100 mg cisplatin per square meter during 2 h in 6 liters of a heating solution (temperature 42 degrees C, flow rate 200 ml/min in a closed circuit) after previous surgical resection. Tumor necrosis factor-alpha (TNF-alpha) and interleukin-6 (IL-6) were measured at 0, 30, 45, 60 and 90 min, both in the blood stream and in the heating solution circulating intraperitoneally. We observed a slight increase in plasma IL-6 occurring as soon as 30 min, a dramatic rise in IL-6 in the heating solution. TNF-alpha values were only slightly augmented. In addition, the importance of various factors in the induction of IL-6 and TNF-alpha production during chemohyperthermia (temperature, mitomycin C, cisplatin) were studied using a whole blood ex vivo model.

Epidermal growth factor receptor-reactive monoclonal antibodies: xenograft antitumor activity alone and as drug immunoconjugates

Antibodies reactive with human epidermal growth factor receptor (EGFr), such as 225 IgG1 (Masui et al., Cancer Res., 44: 1002-1007, 1984), are effective tumor-suppressive agents in xenograft models. In the present study an additional antibody reactive with EGFr was made and compared to 225 IgG1 for antitumor activity as an unmodified antibody or as a drug immunoconjugate. This IgG1 clone, designated EGFrL11, competed with EGF and immunoprecipitated a Mr 178,000 protein identical to that immunoprecipitated with 225 IgG1. Cross-competition and immunodepletion studies indicated that the two antibodies bound to distinct epitopes on the same molecule. Immunofluorescence studies confirmed that the EGFrL11 epitope was expressed on the surface of viable human squamous cell carcinoma lines including T222. Unmodified EGFrL11 and 225 IgG1 were tested for antitumor activity in T222 xenografts. At a dose of 81 mg/kg given twice weekly for 3 weeks, tumor suppression, but not regression, occurred with EGFrL11. A similar result was obtained with 225 IgG1. To gauge the potential of these antibodies as immunoconjugates, both were tested for antitumor activity in the T222 model after conjugation to the Vinca derivative 4-desacetylvinblastine-3-carboxhydrazide. Both immunoconjugates completely regressed established tumors. These data suggest that Vinca conjugates with EGFr-reactive monoclonal antibodies warrant further investigation as possible clinical candidates.

Induction of immunogenicity of monoclonal antibodies by conjugation with drugs

Human anti-mouse antibody has been a nearly consistent result of human clinical trials utilizing murine antibodies. It is generally anticipated that the problem of human anti-mouse antibody will be reduced as genetically engineered, more human ("humanized") antibodies become available. It is not clear, however, what effect chemical modification of such "humanized" antibodies will have on their immunogenicity. The present studies utilize a mouse antibody and rat host model to explore aspects of this question. Rats injected with unmodified mouse monoclonal antibodies failed to mount anti-mouse immune responses, presumably due to their phylogenetic relatedness. In contrast, rats injected with a Vinca immunoconjugate mounted strong anticonjugate antibody responses that were directed primarily against the linker portion of the conjugate. The in vivo serum pharmacokinetics of 125I-labeled antibody and conjugates were evaluated in rats with existing anticonjugate antibody. The peak serum level attained was inversely correlated with the level of reactivity of the anticonjugate antibody with the injected compound. This model provides a potentially useful tool for exploration of the immunogenicity of drug, toxin, or radionuclide monoclonal antibody conjugates.

Antitumor activity of L/1C2-4-desacetylvinblastine-3-carboxhydrazide immunoconjugate in xenografts

The murine IgG3 monoclonal antibody L/1C2 is reactive with a high percentage of human carcinomas and has preferentially strong reactivity with tumors of squamous differentiation. This antibody was tested for antitumor activity in vitro and in xenograft models as a carbohydrate-linked immunoconjugate with the Vinca derivative 4-desacetylvinblastine-3-carboxhydrazide (DAVLBHYD). The conjugate retained good immunoreactivity and was highly active in a cytotoxicity assay. In human tumor nude mouse xenograft studies, L/1C2-DAVLBHYD antitumor activity was superior to that seen with free drug, free antibody, mixtures of free drug and free antibody, or control DAVLBHYD conjugates prepared with non-tumor-binding IgGs. With well-established tumors, potent antitumor activity was observed, including the ability to specifically regress greater than 400-mg tumors to 0 mg. In some cases, apparent long-term cures were effected. In studies using six different human tumor xenografts, the level of potency of L/1C2-DAVLBHYD was related to L/1C2 antigen expression, although the growth rate probably also contributes to the conjugate sensitivity of the tumors.

Monoclonal antibody L/1C2 reactive with a human carcinoma associated antigen

A murine IgG3 monoclonal antibody which defines a human carcinoma associated antigen is described. The antibody, designated L/1C2, was made against a human lung squamous cell carcinoma line designated USCLS-1. It reacts with the surface of 15 of 16 viable human carcinoma cell lines and was detected in 70 of 78 frozen sections of human carcinomas. Melanoma cell lines and frozen sections of melanomas and a lymphoma were unreactive. Normal tissue reactivity included vessels, plus some ducts, glandular structures, and epithelial surfaces. Similar normal tissue reactivity patterns were seen with Rhesus monkey tissue samples. Immunoprecipitation studies indicate that L/1C2 reacts with a glycoprotein doublet which migrates in the range of 110,000 to 140,000 Mr under reducing conditions. Fluorescence analysis suggests this antigen is internalized following reaction with the L/1C2 antibody. Using in vitro human tumor cell growth inhibition assays, it was possible to achieve significant growth inhibition with L/1C2, while another target cell-reactive antibody in the same assay had no inhibitory effect.