Search

PDF file, 75K, Overexpression of CTMP enhances PI3K-mediated Akt phosphorylation.

PDF file, 146K.

PDF file, 187K, The N-terminal domain, 1-64 a.a., of CTMP is required, but not sufficient, to induce Akt phosphorylation, colony formation and cell proliferation.

PDF file, 63K, Association of CTMP expression with prognosis of patients with different cancer type in 4 cohorts from Prognoscan database.

PDF file - 287K, Cisplatin-induced enrichment of CD133+ cells

PDF file - 289K, Carboplatin induced DNA damage and CD133+ cell enrichment

PDF file - 342K, Inhibition of cisplatin-induced enrichment of CD133+ cells by Notch1 knockdown

PDF file, 70K, PI3K inibitor blocked serum-induced Akt phosphorylation.

PDF file - 255K, The role of ABC transporters in cisplatin-induced multi-drug resistance to paclitaxol in H460 and H661 cell lines

PDF file - 143K

PDF file - 2.1MB, Cell differentiation and migration of GFP+ cells isolated from H460 cell line

PDF file, 114K, CTMP mRNA expression data adapted from GEO public available microarray database, GSE3744.

PDF file - 138K, The role of Notch signaling in enrichment of CD133+ cells and multi-drug resistance of H661 cells

Computing-in-memory (CIM) based on SRAM is a promising approach to achieving energy-efficient multiply-and-accumulate (MAC) operations in artificial intelligence (AI) edge devices; however, existing SRAM-CIM chips support only DNN inference. The flow of training data requires that CIM arrays perform convolutional computation using transposed weight matrices. This article presents a two-way transpose (TWT) multiply cell with high resistance to process variation and a novel read scheme that uses input-aware zone prediction of maximum partial MAC values to enhance the signal margin for robust readout. A 28-nm 64-kb TWT CIM macro fabricated using foundry-provided compact 6T-SRAM cells achieved ${T_{AC}}$ of 3.8-21 ns and energy efficiency of 7-61.1 TOPS/W in performing MAC operations using 2-8-b inputs, 4-8-b weights, and 10-20-b outputs.

This article presents a computing-in-memory (CIM) structure aimed at improving the energy efficiency of edge devices running multi-bit multiply-and-accumulate (MAC) operations. The proposed scheme includes a 6T SRAM-based CIM (SRAM-CIM) macro capable of: 1) weight-bitwise MAC (WbwMAC) operations to expand the sensing margin and improve the readout accuracy for high-precision MAC operations; 2) a compact 6T local computing cell to perform multiplication with suppressed sensitivity to process variation; 3) an algorithm-adaptive low MAC-aware readout scheme to improve energy efficiency; 4) a bitline header selection scheme to enlarge signal margin; and 5) a small-offset margin-enhanced sense amplifier for robust read operations against process variation. A fabricated 28-nm 64-kb SRAM-CIM macro achieved access times of 4.1–8.4 ns with energy efficiency of 11.5–68.4 TOPS/W, while performing MAC operations with 4- or 8-b input and weight precision.

Leptomeningeal metastasis (LM) occurs when tumor cells spread to the leptomeningeal space surrounding the brain and the spinal cord, thereby causing poor clinical outcomes. The triple-negative breast cancer (TNBC) has been associated with symptoms of LM and mechanism remained unclear. Through proteomic analysis, we identified high expression of ICAM2 in leptomeningeal metastatic TNBC cells, which promoted the colonization of the spinal cord and resulted in poor survival in vivo. Two-way demonstration indicated that high levels of ICAM2 promoted blood–cerebrospinal fluid barrier (BCB) adhesion, trans-BCB migration, and stemness abilities and determined the specificity of LM in vivo. Furthermore, pulldown and antibody neutralizing assay revealed that ICAM2 determined the specificity of LM through interactions with ICAM1 in the choroid plexus epithelial cells. Therefore, neutralizing ICAM2 can attenuate the progression of LM and prolong survival in vivo. The results suggested that targeting ICAM2 is a potential therapeutic strategy for LM in TNBC.

Currently, the survival rate for breast cancer is more than 90%, but once the cancer cells metastasize to distal organs, the survival rate is dramatically reduced, to less than 30%. Triple-negative breast cancer accounts for 15-20% of all breast cancers. Triple-negative breast cancer (TNBC) is associated with poor prognostic and diagnostic outcomes due to the limiting therapeutic strategies, relative to non-TNBC breast cancers. Therefore, the development of targeted therapy for TNBC metastasis remains an urgent issue. In this study, high Carboxyl-terminal modulator protein (CTMP) is significantly associated with recurrence and disease-free survival rate in TNBC patients. Overexpression of CTMP promotes migration and invasion abilities in BT549 cells. Down-regulating of CTMP expression inhibits migration and invasion abilities in MDA-MB-231 cells. In vivo inoculation of high-CTMP cells enhances distant metastasis in mice. The metastasis incidence rate is decreased in mice injected with CTMP-downregulating MDA-MB-231 cells. Gene expression microarray analysis indicates the Akt-dependent pathway is significantly enhanced in CTMP overexpressing cells compared to the parental cells. Blocking Akt activation via Akt inhibitor treatment or co-expression of the dominant-negative form of Akt proteins successfully abolishes the CTMP mediating invasion in TNBC cells. Our findings suggest that CTMP is a potential diagnostic marker for recurrence and poor disease-free survival in TNBC patients. CTMP promotes TNBC metastasis via the Akt-activation-dependent pathway.

A role for the JAM family in cancer progression has been demonstrated. Nevertheless, there has been no systematic study of the JAM family in specific cancers through clinical specimens, and the mechanism of the molecule is unclear, particularly in subtypes of breast cancer (BRCA). Based on an extensive analysis of the relative expression of the JAM family in cancer by pan-cancer, we found that the increase of F11R (JAM-A) is the opposite of that of JAM-2, or JAM-3, which is unique to BRCA. BRCA prognoses are related to F11R expression, including survival and recurrence-free survival. It is also noteworthy that F11 has different expression characteristics in subtypes, especially in triple-negative breast cancer (TNBC). Using immunohistochemistry (IHC) analysis as an analytical validation, this study provides consistent results with in-silico. As a result of artificial manipulations of F11R combined with microarray-based analyses, we identified that F11R plays an important role in several aspects of cell motility signaling. An upstream regulators and correlation analysis found that EP300 transcription factors may be involved in TNBC metastasis by modulating RHOA, GSK3B, and TGFBR1 expression through F11R-mediated epithelial-mesenchymal transition (EMT). Accordingly, the findings of this study provide a comprehensive interpretation of the relationship between the clinical value of F11R in BRCA subtype progression and the mechanisms that induce metastasis. These results provide a possible target for the treatment of BRCA.

Brian metastasis, which is diagnosed in 30% of triple-negative breast cancer (TNBC) patients with metastasis, causes poor survival outcomes. Growing evidence has characterized miRNAs involving in breast cancer brain metastasis; however, currently, there is a lack of prognostic plasma-based indicator for brain metastasis. In this study, high level of miR-211 can act as brain metastatic prognostic marker in vivo. High miR-211 drives early and specific brain colonization through enhancing trans-blood–brain barrier (BBB) migration, BBB adherence, and stemness properties of tumor cells and causes poor survival in vivo. SOX11 and NGN2 are the downstream targets of miR-211 and negatively regulate miR-211-mediated TNBC brain metastasis in vitro and in vivo. Most importantly, high miR-211 is correlated with poor survival and brain metastasis in TNBC patients. Our findings suggest that miR-211 may be used as an indicator for TNBC brain metastasis.

Here, we found that STK4 was significantly downregulated in colon cancer and was associated with distal metastasis and poor survival. STK4 colocalized with β‐catenin and directly phosphorylated β‐catenin resulting in its degradation via the ubiquitin‐mediated pathway. This may suggest that STK4 knockdown causes β‐catenin phosphorylation failure and subsequently β‐catenin accumulation, consequently leading to anchorage‐independent growth and metastasis in colon cancer., Mammalian STE20‐like kinase 1 (MST1/STK4/KRS2) encodes a serine/threonine kinase that is the mammalian homolog of Drosophila Hippo. STK4 plays an important role in controlling cell growth, apoptosis, and organ size. STK4 has been studied in many cancers with previous studies indicating an involvement in colon cancer lymph node metastasis and highlighting its potential as a diagnostic marker for colon cancer. However, the role of STK4 defect in promoting colon cancer progression is still understudied. Here, we found that STK4 was significantly downregulated in colon cancer and was associated with distal metastasis and poor survival. Furthermore, STK4 knockdown enhanced sphere formation and metastasis in vitro and promoted tumor development in vivo. We found that STK4 colocalized with β‐catenin and directly phosphorylated β‐catenin resulting in its degradation via the ubiquitin‐mediated pathway. This may suggest that STK4 knockdown causes β‐catenin phosphorylation failure and subsequently β‐catenin accumulation, consequently leading to anchorage‐independent growth and metastasis in colon cancer. Our results support that STK4 may act as a potential candidate for the assessment of β‐catenin‐mediated colon cancer prognosis.

Cancer metabolic reprogramming promotes tumorigenesis and metastasis; however, the underlying molecular mechanisms are still being uncovered. In this study, we show that the glycolytic enzyme aldolase A (ALDOA) is a key enzyme involved in lung cancer metabolic reprogramming and metastasis. Overexpression of ALDOA increased migration and invasion of lung cancer cell lines in vitro and formation of metastatic lung cancer foci in vivo. ALDOA promoted metastasis independent of its enzymatic activity. Immunoprecipitation and proteomic analyses revealed γ-actin binds to ALDOA; blocking this interaction using specific peptides decreased metastasis both in vitro and in vivo. Screening of clinically available drugs based on the crystal structure of ALDOA identified raltegravir, an antiretroviral agent that targets HIV integrase, as a pharmacologic inhibitor of ALDOA-γ-actin binding that produced antimetastatic and survival benefits in a xenograft model with no significant toxicity. In summary, ALDOA promotes lung cancer metastasis by interacting with γ-actin. Targeting this interaction provides a new therapeutic strategy to treat lung cancer metastasis. Significance: This study demonstrates the role of aldolase A and its interaction with γ-actin in the metastasis of non–small lung cancer and that blocking this interaction could be an effective cancer treatment.

Epithelial and mesenchymal transition mechanisms continue to occur during the cell cycle and throughout human development from the embryo stage to death. In embryo development, epithelial-mesenchymal transition (EMT) can be divided into three essential steps. First, endoderm, mesoderm, and neural crest cells form, then the cells are subdivided, and finally, cardiac valve formation occurs. After the embryonic period, the human body will be subjected to ongoing mechanical stress or injury. The formation of a wound requires EMT to recruit fibroblasts to generate granulation tissues, repair the wound and re-create an intact skin barrier. However, once cells transform into a malignant tumor, the tumor cells acquire the characteristic of immortality. Local cell growth with no growth inhibition creates a solid tumor. If the tumor cannot obtain enough nutrition in situ, the tumor cells will undergo EMT and invade the basal membrane of nearby blood vessels. The tumor cells are transported through the bloodstream to secondary sites and then begin to form colonies and undergo reverse EMT, the so-called “mesenchymal-epithelial transition (MET).” This dynamic change involves cell morphology, environmental conditions, and external stimuli. Therefore, in this manuscript, the similarities and differences between EMT and MET will be dissected from embryonic development to the stage of cancer metastasis.

Recent SRAM-based computation-in-memory (CIM) macros enable mid-to-high precision multiply-and-accumulate (MAC) operations with improved energy efficiency using ultra-small/small capacity (0.4-8KB) memory devices. However, advanced CIM-based edge-AI chips favor multiple mid/large capacity SRAM-CIM macros: with high input (IN) and weight (W) precision to reduce the frequency of data reloads from external DRAM, and to avoid the need for additional SRAM buffers or ultra-large on-chip weight buffers. However, enlarging memory capacity and throughput increases the delay parasitics on WLs and BLs, and the number of parallel computing elements; resulting in longer compute latency (t AC ), lower energy-efficiency (EF), degraded signal margin, and larger fluctuations in power consumption across data-patterns (see Fig. 16.3.1). Recent SRAM-CIM macros tend to not use in-lab SRAM cells, with a logic-based layout, in favor of foundry provided compact-layout 8T [2], 3, [5] or 6T cells with local-computing cells (LCCs) [4], [6] to reduce the cell-array area and facilitate manufacturing. This paper presents a SRAM-CIM structure using (1) a segmented-BL charge-sharing (SBCS) scheme for MAC operations, with low energy consumption and a consistently high signal margin across MAC values (MACV); (2) An new LCC cell, called a source-injection local-multiplication cell (SILMC), to support the SBCS scheme with a consistent signal margin against transistor process variation; and (3) A prioritized-hybrid-ADC (Ph-ADC) to achieve a small area and power overhead for analog readout. A 28nm 384kb SRAM-CIM macro was fabricated using a foundry compact-6T cell with support for MAC operations with 16 accumulations of 8b-inputs and 8b-weights with near-full precision output (20b). This macro achieves a 7.2ns t AC and a 22.75TOPS/W EF for 8b-MAC operations with an FoM (IN-precision × W-precision × output-ratio × output-channel × EF/t AC ) 6× higher than prior work.

Background Cancer subtype switching, which involves unclear cancer cell origin, cell fate decision, and transdifferentiation of cells within a confined tumor microenvironment, remains a major problem in pancreatic cancer (PDA). Results By analyzing PDA subtypes in The Cancer Genome Atlas, we identified that epigenetic silencing of apoptosis-associated tyrosine kinase (AATK) inversely was correlated with mRNA expression and was enriched in the quasi-mesenchymal cancer subtype. By comparing early mouse pancreatic lesions, the non-invasive regions showed AATK co-expression in cells with acinar-to-ductal metaplasia, nuclear VAV1 localization, and cell cycle suppression; but the invasive lesions conversely revealed diminished AATK expression in those with poorly differentiated histology, cytosolic VAV1 localization, and co-expression of p63 and HNF1α. Transiently activated AATK initiates acinar differentiation into a ductal cell fate to establish apical-basal polarization in acinar-to-ductal metaplasia. Silenced AATK and ectopically expressed p63 and HNF1α allow the proliferation of ductal PanINs in mice. Conclusion Epigenetic silencing of AATK regulates the cellular transdifferentiation, proliferation, and cell cycle progression in converting PDA-subtypes.

Advanced AI edge chips require multibit input (IN), weight (W), and output (OUT) for CNN multiply-and-accumulate (MAC) operations to achieve an inference accuracy that is sufficient for practical applications. Computing-in-memory (CIM) is an attractive approach to improve the energy efficiency $(\mathrm{EF}_{\mathrm{MAC}}]$ of MAC operations under a memory-wall constraint. Previous SRAM-CIM macros demonstrated a binary MAC [4], an in-array 8b W-merging with near-memory computing (NMC) using 6T SRAM cells (limited output precision) [5], a 7b1N-1 bW MAC using a 10T SRAM cell (large area) [3], an 4b1N-5bW MAC with a T8T SRAM cell [1], and 8b1N-1bW NMC with 8T SRAM (long MAC latency $(T_{\mathrm{AC}})$ ) [2]. However, previous works have not achieved high IN/W/OUT precision with fast $\mathrm{T}_{\mathrm{AC}}$ compact-area, high $\mathrm{EF}_{\mathrm{MAC}}$ , and robust readout against process variation, due to (1) small sensing margin in word-wise multiple-bit MAC operations, (2) a tradeoff between read accuracy vs. area overhead under process variation, (3) limited $\mathrm{EF}_{\mathrm{MAC}}$ due to decoupling of software and hardware development.

Many Al edge devices require local intelligence to achieve fast computing time (t AC ), high energy efficiency (EF), and privacy. The transfer-learning approach is a popular solution for Al edge chips, wherein data used to re-train the Al in the cloud is used to fine-tune (re-train) a few of the neural layers in edge devices. This enables the dynamic incorporation of data from in-situ environments or private information. Computing-in-memory (CIM) is a promising approach to improve EF for Al edge chips, existing CIM schemes support inference [1]–[5] with forward (FWD) propagation; however, they do not support training, requiring both FWD and backward (BWD) propagation, due to differences in weight-access flow for FWD and BWD propagation. As Fig. 15.2.1 shows, efforts to increase the precision of the input (IN), weight (W), and/or output (OUT) tend to degrade r AC and EF for training operations irrespective of scheme: digital FWD and BWD (DF-DB) or CIM-FWD-digital-BWD (CiMF-DB). This work develops a two-way transpose (TWT) SRAM-CIM macro supporting multibit MAC operations for FWD and BWD propagation with fast r AC and high EF within a compact area. The proposed scheme features (1) A TWT multiply cell (TWT-MC) with a high resistance to process variation; and (2) a small-offset gain-enhancement sense amplifier (SOGE-SA) to tolerate a small read margin. A 28nm 64Kb TWT SRAM-CIM macro was fabricated using a foundry-provided compact 6T-SRAM cell for SRAM-CIM devices supporting both inference and training operations for the first time. This macro also demonstrates the fastest t AC (3.8 – 21ns) and highest EF (7 – 61.1TOPS/w) for MAC operations using 2 – 8b inputs, 4 – 8b weights and 12 − 20b outputs.

Alzheimer's disease (AD) is characterized by a chronic decline in cognitive function and is pathologically typified by cerebral deposition of amyloid-β peptide (Aβ). The production of Aβ is mediated by sequential proteolysis of amyloid precursor protein (APP) by β- and γ-secretases, and has been implicated as the essential determinant of AD pathology. Previous studies have demonstrated that the level of phosphatidylinositol-4,5-bisphosphate [PI(4,5)P2] in the membrane may potentially modulate Aβ production. Given that PI(4,5)P2 is produced by type 1 phosphatidylinositol-4-phosphate 5-kinases (PIP5Ks), we sought to determine whether the level of PIP5K type Iα (PIP5K1A) can affect production of Aβ by modulating the lipid composition of the membrane. Using a HEK-derived cell line that constitutively expresses yellow fluorescent protein-tagged APP (APP-YFP), we demonstrated that overexpression of PIP5K1A results in significant enhancement of non-amyloidogenic APP processing and a concomitant suppression of the amyloidogenic pathway, leading to a marked decrease in secreted Aβ. Consistently, cells overexpressing PIP5K1A exhibited a significant redistribution of APP-YFP from endosomal compartments to the cell surface. Our findings suggest that PIP5K1A may play a critical role in governing Aβ production by modulating membrane distribution of APP, and as such, the pathway may be a valuable therapeutic target for AD.