Secure State Estimation for Artificial Neural Networks With Unknown-But-Bounded Noises: A Homomorphic Encryption Scheme

This article is concerned with the secure state estimation problem for artificial neural networks (ANNs) subject to unknown-but-bounded noises, where sensors and the remote estimator are connected via open and bandwidth-limited communication networks. Using the encoding-decoding mechanism (EDM) and the Paillier encryption technique, a novel homomorphic encryption scheme (HES) is introduced, which aims to ensure the secure transmission of measurement information within communication networks that are constrained by bandwidth. Under this encoding-decoding-based HES, the data being transmitted can be encrypted into ciphertexts comprising finite bits. The emphasis of this research is placed on the development of a secure set-membership state estimation algorithm, which allows for the computation of estimates using encrypted data without the need for decryption, thereby ensuring data security throughout the entire estimation process. Taking into account the unknown-but-bounded noises, the underlying ANN, and the adopted HES, sufficient conditions are determined for the existence of the desired ellipsoidal set. The related secure state estimator gains are then derived by addressing optimization problems using the Lagrange multiplier method. Lastly, an example is presented to verify the effectiveness of the proposed secure state estimation approach.

[Clinical efficacy of induction chemoimmunotherapy for locally advanced hypopharyngeal carcinoma: a prospective phase Ⅱ study]

Objective: To evaluate the objective response rate (ORR) of induction chemoimmunotherapy with camrelizumab plus TPF (docetaxel, cisplatin, and capecitabine) for locally advanced hypopharyngeal squamous cell carcinoma (LA HSCC) and potential predictive factors for ORR. Methods: A single-center, prospective, phase 2 and single-arm trial was conducted for evaluating antitumor activity of camrelizumab+TPF(docetaxel+cisplatin+capecitabine) for LA HSCC between May 21, 2021 and April 15, 2023, patients admitted to the Eye & ENT Hospital affiliated with Fudan University. The primary endpoint was ORR, and enrolled patients with LA HSCC at T3-4N0-3M0 received induction chemoimmunotherapy for three cycles: camrelizumab 200 mg day 1, docetaxel 75 mg/m2 day 1, cisplatin 25 mg/m2 days 1-3, and capecitabine 800 mg/m2 days 1-14. Patients were assigned to radioimmunotherapy when they had complete response or partial response (PR)>70% (Group A), or assigned to surgery plus adjuvant radiotherapy/chemoradiotherapy when they had PR≤70% (Group B), and the responses were defined by using tumor volume evaluation system. Tumor diameter was also used to assess the treatment responses by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Use SPSS 23.0 software was used to analyze the data. Results: A total of 51 patients were enrolled who underwent the induced chemoimmunotherapy for three cycles, and all were males, aged 35-69 years old. After three cycles of induction immunochemotherapy, 42 (82.4%) patients existed in Group A (complete response or PR>70%) and 9 patients (17.6%) in Group B (PR≤70%), the ORR was 82.4%. The primary endpoint achieved expected main research objectives. Compared to the patients of Group A, the patients of Group B showed the higher T stage and the larger volume of primary tumor before induced immunochemotherapy, and also had the less regression of tumor volume after induced immunochemotherapy (all P<0.05). The optimal cutoff value of pre-treatment tumor volume for predicting ORR was 39 cm3. The T stage (OR=12.71, 95%CI: 1.4-112.5, P=0.022) and the volume (OR=7.1, 95%CI: 1.4-36.8, P=0.018) of primary tumor were the two main factors affecting ORR rate of induction chemoimmunotherapy. Conclusion: The induction chemoimmunotherapy with camrelizumab plus TPF shows an encouraging antitumor efficacy in LA HSCC.

GSB: Group superposition binarization for vision transformer with limited training samples

Vision Transformer (ViT) has performed remarkably in various computer vision tasks. Nonetheless, affected by the massive amount of parameters, ViT usually suffers from serious overfitting problems with a relatively limited number of training samples. In addition, ViT generally demands heavy computing resources, which limit its deployment on resource-constrained devices. As a type of model-compression method, model binarization is potentially a good choice to solve the above problems. Compared with the full-precision one, the model with the binarization method replaces complex tensor multiplication with simple bit-wise binary operations and represents full-precision model parameters and activations with only 1-bit ones, which potentially solves the problem of model size and computational complexity, respectively. In this paper, we investigate a binarized ViT model. Empirically, we observe that the existing binarization technology designed for Convolutional Neural Networks (CNN) cannot migrate well to a ViT's binarization task. We also find that the decline of the accuracy of the binary ViT model is mainly due to the information loss of the Attention module and the Value vector. Therefore, we propose a novel model binarization technique, called Group Superposition Binarization (GSB), to deal with these issues. Furthermore, in order to further improve the performance of the binarization model, we have investigated the gradient calculation procedure in the binarization process and derived more proper gradient calculation equations for GSB to reduce the influence of gradient mismatch. Then, the knowledge distillation technique is introduced to alleviate the performance degradation caused by model binarization. Analytically, model binarization can limit the parameter's search space during parameter updates while training a model. Therefore, the binarization process can actually play an implicit regularization role and help solve the problem of overfitting in the case of insufficient training data. Experiments on three datasets with limited numbers of training samples demonstrate that the proposed GSB model achieves state-of-the-art performance among the binary quantization schemes and exceeds its full-precision counterpart on some indicators. Code and models are available at: https://github.com/IMRL/GSB-Vision-Transformer.

Hamiltonian-Driven Adaptive Dynamic Programming With Efficient Experience Replay

This article presents a novel efficient experience-replay-based adaptive dynamic programming (ADP) for the optimal control problem of a class of nonlinear dynamical systems within the Hamiltonian-driven framework. The quasi-Hamiltonian is presented for the policy evaluation problem with an admissible policy. With the quasi-Hamiltonian, a novel composite critic learning mechanism is developed to combine the instantaneous data with the historical data. In addition, the pseudo-Hamiltonian is defined to deal with the performance optimization problem. Based on the pseudo-Hamiltonian, the conventional Hamilton-Jacobi-Bellman (HJB) equation can be represented in a filtered form, which can be implemented online. Theoretical analysis is investigated in terms of the convergence of the adaptive critic design and the stability of the closed-loop systems, where parameter convergence can be achieved under a weakened excitation condition. Simulation studies are investigated to verify the efficacy of the presented design scheme.

LAFIT: Efficient and Reliable Evaluation of Adversarial Defenses With Latent Features

Deep convolutional neural networks (CNNs) can be easily tricked to give incorrect outputs by adding tiny perturbations to the input that are imperceptible to humans. This makes them susceptible to adversarial attacks, and poses significant security risks to deep learning systems, and presents a great challenge in making CNNs robust against such attacks. An influx of defense strategies have thus been proposed to improve the robustness of CNNs. Current attack methods, however, may fail to accurately or efficiently evaluate the robustness of defending models. In this paper, we thus propose a unified lp white-box attack strategy, LAFIT, to harness the defender's latent features in its gradient descent steps, and further employ a new loss function to normalize logits to overcome floating-point-based gradient masking. We show that not only is it more efficient, but it is also a stronger adversary than the current state-of-the-art when examined across a wide range of defense mechanisms. This suggests that adversarial attacks/defenses could be contingent on the effective use of the defender's hidden components, and robustness evaluation should no longer view models holistically.

Improving Pretrained Language Model Fine-Tuning With Noise Stability Regularization

The advent of large-scale pretrained language models (PLMs) has contributed greatly to the progress in natural language processing (NLP). Despite its recent success and wide adoption, fine-tuning a PLM often suffers from overfitting, which leads to poor generalizability due to the extremely high complexity of the model and the limited training samples from downstream tasks. To address this problem, we propose a novel and effective fine-tuning framework, named layerwise noise stability regularization (LNSR). Specifically, our method perturbs the input of neural networks with the standard Gaussian or in-manifold noise in the representation space and regularizes each layer's output of the language model. We provide theoretical and experimental analyses to prove the effectiveness of our method. The empirical results show that our proposed method outperforms several state-of-the-art algorithms, such as [Formula: see text] norm and start point (L2-SP), Mixout, FreeLB, and smoothness inducing adversarial regularization and Bregman proximal point optimization (SMART). In addition to evaluating the proposed method on relatively simple text classification tasks, similar to the prior works, we further evaluate the effectiveness of our method on more challenging question-answering (QA) tasks. These tasks present a higher level of difficulty, and they provide a larger amount of training examples for tuning a well-generalized model. Furthermore, the empirical results indicate that our proposed method can improve the ability of language models to domain generalization.

Hamiltonian-Driven Adaptive Dynamic Programming With Approximation Errors

In this article, we consider an iterative adaptive dynamic programming (ADP) algorithm within the Hamiltonian-driven framework to solve the Hamilton-Jacobi-Bellman (HJB) equation for the infinite-horizon optimal control problem in continuous time for nonlinear systems. First, a novel function, "min-Hamiltonian," is defined to capture the fundamental properties of the classical Hamiltonian. It is shown that both the HJB equation and the policy iteration (PI) algorithm can be formulated in terms of the min-Hamiltonian within the Hamiltonian-driven framework. Moreover, we develop an iterative ADP algorithm that takes into consideration the approximation errors during the policy evaluation step. We then derive a sufficient condition on the iterative value gradient to guarantee closed-loop stability of the equilibrium point as well as convergence to the optimal value. A model-free extension based on an off-policy reinforcement learning (RL) technique is also provided. Finally, numerical results illustrate the efficacy of the proposed framework.

[Relationship of serum uric acid with prediabetes and newly detected type 2 diabetes mellitus]

Objective: To evaluate the relationship of serum uric acid with prediabetes and newly detected type 2 diabetes mellitus (T2DM) in adults. Methods: Data were obtained from the baseline investigation of Songjiang Peak-Plan cohort. According to the baseline fasting plasma glucose and glycosylated hemoglobin, the eligible subjects were divided into normal blood sugar group, prediabetes group, and newly detected T2DM group. Unconditional logistic regression model was used to explore the effect of serum uric acid level on prediabetes and newly detected T2DM, and restricted cubic spline (RCS) function was used to explore the nonlinear dose-response relationship of serum uric acid level with the prevalence of prediabetes and newly detected T2DM. Results: A total of 30 375 subjects were included in the analysis, with an average age of (55.36±11.52) years, and 60.2% (18 299) of them were women. The baseline survey found that the prevalence of prediabetes was 38.6% (11 739 cases), and the prevalence of newly detected T2DM was 6.6% (1 992 cases). Logistic regression analysis showed that, in women, for every 10µmol/L increase in serum uric acid, the risk of developing prediabetes and T2DM s increased by 2.4% (OR=1.024, 95%CI: 1.018-1.030), and 1.5% (OR=1.015, 95%CI: 1.005-1.025), respectively; in men, for every 10 µmol/L increase in serum uric acid, the risk of developing prediabetes and T2DM decreased by 0.8% (OR=0.992, 95%CI: 0.987-0.998) and 5.0% (OR=0.950, 95%CI: 0.939-0.960), respectively. The RCS function showed that the serum uric acid level showed a nonlinear dose-response relationship with newly detected T2DM (P=0.017), but not with prediabetes (P=0.670) in women and showed a nonlinear dose-response relationship with both prediabetes (P=0.040) and newly detected T2DM (P<0.001) in men. Conclusions: Adult women are at increased risk of prediabetes and newly detected T2DM with increase of serum uric acid level, and adult men are at decreased risk of newly diagnosed T2DM with the increase of serum uric acid level. There was no significant relationship between serum uric acid level and prediabetes in men.

[Feasibility and perioperative safety of transoral robotic surgery with da Vinci Xi platform]

Objective: To explore the feasibility and perioperative safety of transoral robotic surgery with da Vinci Xi platform for pharyngolaryngeal tumors. Methods: A retrospective analysis was performed on 55 consecutive cases with resection of pharyngolaryngeal tumors by transoral robotic surgery with da Vinci Xi platform from July 27, 2020 to October 31, 2021 in the Department of Head and Neck Surgery, Fudan University Eye, Ear, Nose and Throat Hospital, including 44 males and 11 females, aged 25-79 years. There were 41 cases of oropharyngeal tumors, 9 cases of parapharyngeal space tumors, 2 cases of laryngeal tumors, 2 cases of hypopharyngeal tumors and 1 case of retropharyngeal space tumor. Operative time, intraoperative blood loss, postoperative hospital stay, perioperative tracheotomy, nasal feeding, hemorrhage and other complications were analyzed. Results: Of the 55 patients, 54 received resection of pharyngolaryngeal tumors by da Vinci robot through oral approach, and only 1 case of pyriform sinus carcinoma underwent a conversion to open surgery due to poor exposure of lower margin. The average surgical time for the patients with transoral robotic surgeries was 64.4 min, the average blood loss was 24.8 ml, the average postoperative hospital stay was 6.9 d, and the average oral feeding time was 11.1 d. Seventeen patients (30.9%) underwent preventive tracheotomy during surgery. Among 38 cases of laryngeal cancer, 28 underwent simultaneously neck dissection. No serious complications occurred in all patients during and after operation. The follow-up time was 1-15 months. Aside from 1 patient had a relapse 10 months after surgery, other patients had no recurrence or metastasis. Conclusion: Transoral robotic surgery with da Vinci Xi is safe, effective and minimally invasive for resection of pharyngolaryngeal tumors under reasonable indications.

[Study on the incidence of adult herpes zoster in Yichang city and its association with early-life famine exposure]

Based on Yichang health big data platform, 850 608 patients from September 2018 to September 2019 were included in this study. According to the date of birth, the participants were divided into early childhood famine exposure group, fetal famine exposure group and non-famine exposure group. The incidence of adult herpes zoster (HZ) in Yichang city was analyzed, and the correlation between early life famine exposure and adult HZ was analyzed. In 2019, the crude incidence rate of adult HZ in Yichang was 6.83‰. The crude incidence rate of adult HZ in females (7.26‰) was higher than that in males (6.40‰). Compared with the non-famine exposure group, fetal famine exposure was associated with the incidence of adult HZ (OR=1.21; 95%CI: 1.01-1.45, P=0.041). After stratification by sex, fetal famine exposure was only found to be associated with the onset of adult HZ in females (OR=1.28, 95%CI:1.02-1.61, P=0.034).

“In-Network Ensemble”: Deep Ensemble Learning with Diversified Knowledge Distillation

Ensemble learning is a widely used technique to train deep convolutional neural networks (CNNs) for improved robustness and accuracy. While existing algorithms usually first train multiple diversified networks and then assemble these networks as an aggregated classifier, we propose a novel learning paradigm, namely, “In-Network Ensemble” ( INE ) that incorporates the diversity of multiple models through training a SINGLE deep neural network. Specifically, INE segments the outputs of the CNN into multiple independent classifiers, where each classifier is further fine-tuned with better accuracy through a so-called diversified knowledge distillation process . We then aggregate the fine-tuned independent classifiers using an Averaging-and-Softmax operator to obtain the final ensemble classifier. Note that, in the supervised learning settings, INE starts the CNN training from random, while, under the transfer learning settings, it also could start with a pre-trained model to incorporate the knowledge learned from additional datasets. Extensive experiments have been done using eight large-scale real-world datasets, including CIFAR, ImageNet, and Stanford Cars, among others, as well as common deep network architectures such as VGG, ResNet, and Wide ResNet. We have evaluated the method under two tasks: supervised learning and transfer learning. The results show that INE outperforms the state-of-the-art algorithms for deep ensemble learning with improved accuracy.

[Pilot surveillance and evaluation of influenza-like illness based on automatic computer analysis of electronic medical record in sentinel hospital]

Objective: To compare the accuracy of influenza-like illness (ILI) surveillance by automatic computer analysis based on electronic medical records and by doctor's report. Methods: A total of 3 542 patients who presented to Yichang Central Hospital fever clinic, respiratory department or emergency department between April to October 2019 with an ICD-10 code for acute respiratory illness (J00-J22) and complete electronic medical information of ILI related syndromes were drawn as the study sample. Taking the classification of the study sample according to the ILI case definition by influenza surveillance professionals as the gold standard, draw the receiver operating characteristic (ROC) curve and calculate sensitivity, specificity, diagnostic consistency to compared the accuracy of ILI surveillance by automatic computer analysis and by doctor's report. Results: Median age of 3 542 cases was 30 (24, 38) years old; 1 179 cases (33.29%) compliance with the case definition, ILI reported by doctors was 1 306 cases (36.87%), and computer automatic identification ILI were 1 150 cases (32.47%); 1 391 (39.27%) cases were men. The results of automatic computer analysis and doctor report consistency of kappa values with gold standard judgment were 0.97 and 0.66 respectively; area under the ROC curve was 0.98 and 0.84, respectively. And the sensitivity and specificity of automatic computer analysis were higher than that of doctor's report (all P values were <0.001), the sensitivity was 96.95% and 82.27%, and the specificity was 99.70% and 85.78%, respectively. Conclusion: The automatic computer analysis based on electronic medical records can identified ILI cases with good sensitivity and specificity in ILI case surveillance.

[Influence of dietary patterns on type 2 diabetes mellitus in local residents aged 40 years and above in Songjiang district, Shanghai]

Objective: To evaluate the association between dietary pattern and type 2 diabetes mellitus (T2DM) in local residents aged 40 years and above in Songjiang district of Shanghai. Methods: Data was obtained from the baseline investigation of Shanghai Peak-Plan cohort. According to inclusion and exclusion criteria, we selected people with T2DM as the case group, and people without T2DM as controls, matched with gender, age and alcohol intake status. Dietary patterns were established by factor analysis. Conditional logistic regression model (CLRM) was used to explore the relationship between different dietary patterns and T2DM, as well as the association between multiplicative interactions of dietary patterns and T2DM. Results: We used factor analysis to obtain six dietary patterns: including meat, desserts-coarse cereals, condiment-egg, beverage, cereals-tubers and fruit-vegetable. Data from multivariate condition logistic regression suggested that condiment-egg patterns as (OR=0.543, 95%CI: 0.377-0.781), beverage (OR=0.590, 95%CI: 0.409-0.852), cereals-tubers (OR=0.592, 95%CI: 0.414-0.848), fruit-vegetable (OR=0.604, 95%CI: 0.417-0.876) were associated with the reduced risks for T2DM. After analyzing the multiplicative interactions between dietary patterns, there were three interaction items associated with T2DM with statistical significances: the multiplicative interaction between meat pattern and condiments-egg pattern was related with the increased risk for T2MD, and the multiplicative interactions between cereal-tubers pattern, meat pattern between cereal-tubers pattern and fruit-vegetable pattern were related with the reduced risks for T2MD. Conclusions: In the six dietary pattern under study, cereals-tubers pattern undwr studly, fruit-vegetable pattern might be more practical in preventing T2DM. In addition, the dietary pattern with cereals and potatos as the main ingredients, appropriate meat and reduced condiment intake might also play positive roles in reducing the risk for T2DM.

[Construction and use of big data for health management]

Population-based lifetime health services rely on health management practice. Collection, management and analysis of big data are highly suitable for the huge population base in China. Nowadays, more and more research focus on the methods, security and ethnicity of health management and big data, and a plenty of instructive results have been made, which could be used to guide the future practice and development. Yichang mode has set a precedent for construction and use of big data for health management.

High calpain-1 expression predicts a poor clinical outcome and contributes to tumor progression in pancreatic cancer patients

BACKGROUND

Pancreatic cancer (PC) is a highly aggressive and metastatic disease, with an elevated mortality rate. It is, therefore, crucial to assess factors affecting the prognosis of PC patients. Meanwhile, calpain-1 is associated with malignant tumor progression and metastasis. Thus, it is meaningful to evaluate the relationship between calpain-1 and PC.

MATERIALS AND METHODS

Calpain-1 protein expression was assessed by immunohistochemistry in 96 pancreatic cancer samples and paired adjacent non-cancerous specimens. In addition, calpain-1 protein levels were assessed in six PC cell lines by western blot (WB). Next, PC cells were transfected with calpain-1 siRNA, and silencing was confirmed by WB. Finally, cell proliferation, colony formation, migration and invasion assays, and cell apoptosis analysis were performed to examine the effects of calpain-1 knockdown on proliferation, growth, apoptosis, migration, and invasion in PC cells.

RESULTS

The results showed that calpain-1 was overexpressed in PC tissues and cells. Meanwhile, calpain-1 overexpression was associated with tumor site (P = 0.029), metastasis (P = 0.000), and TNM stage (P = 0.000), but showed no associations with histological grade (P = 0.396), age (P = 0.809), sex (P = 1.000), and lesion size (P = 0.679). The Kaplan-Meier method demonstrated that the low calpain-1 expression group had increased overall survival (OS) compared with patients expressing high calpain-1 levels (28.7 ± 4.1 vs. 17.0 ± 2.3 months) (P = 0.005). Besides, calpain-1 in PC cells was successfully silenced by liposome-mediated RNA interference, resulting in reduced cell growth, invasion, and metastasis in PC cells, with no effect on apoptosis.

CONCLUSION

The above findings suggest that calpain-1 should be considered a potential biomarker for PC prognosis and therapy.

Effects of feeding lutein on production performance, antioxidative status, and milk quality of high-yielding dairy cows

This experiment was conducted to determine the influences of supplementing different levels of an additive containing lutein in the diet of Chinese Holstein lactating cows on production performance, antioxidative plasma metabolites, and milk quality. This study was performed on 60 multiparous Holstein dairy cows in peak lactation. The cows were randomly allocated to 1 of 4 homogeneous treatments, with lutein preparation (extracted from marigolds; effective lutein content was 2%) added at levels of 0, 100, 150, and 200 g/d per head, with the actual available amounts being 0, 2, 3, and 4 g of lutein/d per head, respectively. The experiment lasted for 13 wk, with the first week for adaptation. Milk yield and milk compositions were recorded weekly, and milk concentrations of lutein, dry matter intake, and antioxidative blood index were analyzed in the first, fourth, seventh, and thirteenth week of the study. The results showed that adding lutein in the diet had no effect on dry matter intake compared with the control group; however, it slowed down the trend of decline in milk yield, and had a linear incremental effect on milk yield with increasing concentration of lutein. Dietary lutein tended to quadratically increase the percentage of milk fat, and linearly increased milk lactose concentration, with the highest value when treated at 200 g of lutein preparation/d per head, and decreased somatic cell count, with the lowest values when treated with 150 and 200 g of lutein preparation/d per head. The concentration of lutein in milk linearly increased with the incorporation of the additive, with a value of 0.59, 0.70, 1.20, and 1.50 μg/100mL when treated with 0, 100, 150, and 200 g/d, respectively. Total plasma antioxidant capacity tended to linearly increase in cows fed lutein preparation, whereas plasma superoxide dismutase and glutathione peroxidase activities did not differ significantly. In conclusion, addition of lutein in the diet could improve the production performance and health status of dairy cows.

Outcomes of laryngectomy in elderly patients with laryngeal carcinoma

Laryngeal carcinoma is a common tumor of the head and neck region. This study aimed to examine the outcomes of laryngectomy in elderly patients with laryngeal carcinoma. One-hundred twenty-two patients (male, 117; female, 5) aged 60 years or older (range, 60-94 years) who underwent laryngectomy between 1996 and 2010 were included. All patients were diagnosed with squamous cell carcinoma of the larynx, and 95 patients (77.9%) had additional concurrent diseases. Tumors were staged according to the TNM categories of the American Joint Committee on Cancer 2002 criteria; there were 16 stage-I, 24 stage-II, 52 stage-III, and 30 stage-IV cases. With regard to treatment modalities, 10 patients underwent transoral laser laryngectomy, 25 underwent partial laryngectomy, and 87 underwent total laryngectomy. When necessary, neck dissection was performed according to the Dalian criteria set in 2004 (a Chinese standard). Of the 122 cases, there were 114 cases of grade I (93.4%), 5 cases of grade II, and 3 cases of grade III (pharyngeal fistula in 2 cases recovered after 2 weeks of care) wound healing. No significant differences were observed in the occurrence or severity of comorbidities. The 1-, 3-, and 5-year actuarial survival rates were 97.5% (119/122), 84.4% (92/109), and 68.4% (67/98), respectively. Age alone should not be used to determine treatment options for elderly patients with squamous cell carcinoma. Presuming that careful pre-treatment evaluations are performed, laryngectomy is a key method for elderly patients with laryngeal carcinoma.

Properties of microtubules assembled from mammalian tubulin synthesized in Escherichia coli

When isolated from tissues, the alpha beta-dimeric protein tubulin consists of multiple isoforms which originate from the expression and subsequent posttranslational modification of multiple polypeptide sequences. Microtubules studied in vitro consist of mixtures of these isoforms. It is therefore not known whether dimers composed of single sequences of alpha- and beta-tubulin can polymerize to form microtubules, or whether posttranslational modifications may be necessary for microtubule assembly. To initiate investigation of these questions, rabbit reticulocyte lysate, which contains the cytoplasmic chaperonin CCT and its cofactors, was employed to prepare substantial quantities (tens of micrograms) of active tubulin by in vitro folding of mouse alpha- and beta-tubulins recombinantly synthesized in E. coli. This recombinant tubulin is composed of only a single alpha-chain and a single beta-chain. When analyzed after folding by isoelectric focusing, each chain yielded only one band, indicating that neither was detectably posttranslationally modified in the course of the folding reaction. When subjected to assembly-promoting conditions, this tubulin formed microtubules without the addition of any exogenous protein. Electron microscopy showed them to be of normal morphology. Analysis of their protein composition showed that they are composed nearly entirely of recombinant tubulin. These results demonstrate that the naturally occurring mixtures of isoforms are not strictly required for the formation of microtubules. They also open a route to other studies, both biomedical and structural, of fully defined tubulin in vitro.

Serum effect on cellular uptake of spermidine, spergualin, 15-deoxyspergualin, and their metabolites by L5178Y cells

Spergualin (SG) and 15-deoxyspergualin (DSG) were more slowly incorporated into L5178Y cells than spermidine. SG and DSG inhibited carrier-mediated transport of [3H]spermidine competitively with inhibition constants of 0.67 mM and 0.45 mM, respectively. Addition of calf serum stimulated uptake of [3H]spermidine into the cells in a serum concentration-dependent manner. The effect was not observed when horse serum was used in place of calf serum. Preincubation of spermidine in calf serum for 1 hour before addition to cells remarkably decreased cellular incorporation of tritium. Three amine oxidase inhibitors, aminoguanidine, 3-hydroxybenzyloxyamine, and semicarbazide, inhibited stimulation of uptake of [3H]spermidine by calf serum and the decrease of it by preincubation in calf serum. So we propose that cellular incorporation or binding of products generated by oxidation of spermidine by amine oxidase in calf serum was much faster than that of spermidine itself and they were unstable and transformed quickly to unincorporable or non-binding substances if cellular targets were not present. Effect of amine oxidase inhibitors on cytotoxic activity of SG and DSG were determined in low and high concentrations of calf serum. In the presence of 10% calf serum in the basal medium, cytotoxicity to L5178Y cells by SG and DSG was suppressed at high drug concentrations (above 10 micrograms/ml) and enhanced at low drug concentrations (below 2.5 micrograms/ml) by amine oxidase inhibitors. In the presence of 0.5% calf serum suppression of cytotoxicity at high drug concentrations by amine oxidase inhibitors was also observed, but enhancement at low drug concentrations was obscure. These data may suggest the existence of two kinds of cytotoxic mechanism of SG and DSG, one dependent on and one independent of amine oxidase in serum.

Cellular uptake and efflux and cytostatic activity of 4'-O-tetrahydropyranyladriamycin in adriamycin-sensitive and resistant tumor cell lines

Cellular uptake and cytostatic activity of 4'-O-tetrahydropyranyladriamycin (THP) in various sublines resistant to anthracycline antibiotics of mouse lymphoblastoma L5178Y, Chinese hamster ovary (CHO) and mouse leukemia P388 cells were studied. All the sublines resistant to adriamycin (ADR) showed slightly decreased uptake of THP as compared with each sensitive lines, but THP was still taken up much more quickly than ADR by each of the ADR-resistant cell lines. Efflux of both anthracycline glycosides from the ADR-resistant P388 cells was faster than that from the ADR-sensitive P388 cells. The percentage of THP retained at equilibrium was higher than that of ADR in both ADR-resistant and -sensitive P388 cells. Cytotoxicity of THP to ADR-resistant cell lines was considerably lower compared with that for each of the sensitive lines but THP inhibited growth of ADR-resistant tumor cells at a concentration about 10 times lower than that for ADR. Thus THP was taken up more quickly, effluxed more slowly than ADR from the ADR-resistant cells, and showed stronger cytostatic activity than ADR on the cells.