HOUSTON--(BUSINESS WIRE)--Kiromic Biopharma, Inc. (Nasdaq: KRBP) is an immuno-oncology target discovery and gene-editing company with a proprietary artificial intelligence neural network platform (Diamond AI) that is used to develop novel oncology therapeutics.
Kiromic is pleased to announce the identification of new targets for the treatment of solid tumors, identified through the application of its bioinformatics platform (CancerDiff and Diamond AI). These new targets include novel epitopes of NY-ESO-1 and a Cancer Selective Isoform Splice Variant of Mesothelin (Isomesothelin). Strikingly, in vitro, and in vivo experiments have demonstrated that the identified IsoMesothelin variant is an excellent target for specific and potent allogeneic CAR Gamma Delta T-cell therapy for solid tumors.
These preclinical results represent a key milestone in the development process of Kiromic and were showcased in 3 posters presented during the annual April 2021 American Association of Cancer Research meeting:
AACR Posters:
|
Identification of an Ovarian Cancer Selective Splice Variant of Mesothelin Utilizing the Kiromic Proprietary Search Engine CancerDiff |
|
1534
|
Mesothelin Isoform 2 is a Novel Target for Allogeneic CAR Gamma Delta T-cell Therapy in Solid Tumors |
|
243
|
Identification of Novel Epitopes of NY-ESO-1 for Solid Malignancies by Kiromic Proprietary Search Engine Diamond |
Poster 247: Showcases how CancerDiff was validated and then used to predict the upregulation of IsoMSLN in ovarian cancer.
Technical
|
CancerDiff was able to: -- identify and validate a unique peptide isoform of mesothelin post transcription within 71% of ovarian cancer (OV) specimens, and -- confirmed the upregulation of IsoMSLN in Ovarian cancer. |
What makes CancerDiff Special vs. nearest competition |
These data sets derived from publicly available proteomic repositories were searched for their unique signature peptide variants that were expressed on the surface of cancer cells. These datasets, from the clinical study “S038 Confirmatory Study performed at Johns Hopkins University”, were downloaded from CPTAC (https://cptac-data-portal.georgetown.edu/study-summary/S038). Data originated from TMT10plex quantification for global proteomic profiling acquired with Orbitrap Fusion Lumos mass spectrometer. 13 datasets were analyzed, comprising 94 ovarian tumor and 23 ovarian normal tissue samples from the same group of ovarian cancer patients. For data parsing and data quality control, data was processed by MS Biowork through the MaxQuant software v1.6.2.3 (www.maxquant.org) for recalibration of MS data, filtering of database search results at the 1% protein and peptide false discovery rate (FDR), calculation of reporter ion intensities (TMT), and isotopic correction of reporter ion intensities (TMT). The identification of unique peptides such as the isoform of mesothelin is allowing us to specifically improve on these unique biomarkers. |
Human labor required to yield equivalent results |
Normally, this requires several scientists working for months to reach the same research conclusion. |
Poster 1534: In collaboration with premier Institutions such as MD Anderson Cancer Center (MDACC) and Baylor Medical College we have validated our first target IsoMSLN and its monoclonal antibody specific binder expressed in indication-specific solid malignancies.
Technical
|
Using a bioinformatics platform (CancerDiff), we searched and analyzed for public gene expression databases
|
What makes CancerDiff Special vs. nearest competition |
In our data mining, we were able to identify cancer-associated antigens caused by alternative splicing. One particular alternative splicing of interest is the isoform of mesothelin. Isoform 1 is the predominant transcript detected in normal and tumor tissues and has been a promising target for cancer immunotherapy in the past. However, Isoform 2 is a minor transcript using alternatively spliced exons producing 8 additional amino acids compared to isoform 1. For example, our data shows that isoform 2 of Mesothelin is selectively expressed in ovarian cancers but not in normal tissues. This alternatively spliced isoform thus makes it an ideal target for incorporation into CAR-T cells. In the past, this process to identify alternatively spliced forms of biomarkers would have taken multiple scientists several months to achieve the same results. This poster shows for the first time the validation of the first target: An IsoMSLN monoclonal antibody specific binder for indication-specific solid malignancies such as ovarian cancer, pancreatic adenocarcinoma, and primary mesothelioma, validating the Anti-IsoMSLN CAR GDT cells as a specific and potent off-the-shelf tumor therapy. (GDT: Gamma Delta T-cells) |
Human labor required to yield equivalent results |
Normally, this requires several scientists working for months to reach the same research conclusion. |
Poster 243: As a validation of our Diamond AI algorithms, we used our Diamond AI algorithm to predict a well-known target like NY-ESO-1.
Technical
|
Using our proprietary software DIAMOND, Kiromic was able to: -- Perform a metanalyses of expression data -- Predict the immunogenicity of NY-ESO-1 peptides Diamond AI’s findings were compared to published expression and immunogenicity data. |
What makes this validation special |
This poster validates the predictive value of DIAMOND algorithms, the meta-analyses of expression data of cancer-testis antigen New York Esophageal Squamous Cell Carcinoma 1 (NY-ESO-1) and predicts immunogenic peptides compared to experimental data in the literature. In agreement with published clinical observations, DIAMOND metanalysis showed NY-ESO-1 genic overexpression in cutaneous melanoma, lung adenocarcinoma, and sarcoma. Taken together these data support DIAMOND as a reliable platform for the discovery of new immunogenic targets for cancer therapy. |
Human labor required to yield equivalent results |
In the past, this calculation of peptide binding affinities to HLA molecules would have taken multiple scientists several months to achieve the same results. |
CEO of Kiromic, Dr. Maurizio Chiriva-Internati, DBSc, PhDs, stated:
“I am very proud to present an Artificial intelligence (AI) platform that can empower scientists and investors to directly leverage data and experiments for generating new ideas and formulate hypotheses.
This is the focus of Kiromic BioPharma (NASDAQ: KRBP), a pre-clinical stage biotechnology company that uses ‘Pragmatic creativity.’
Our novel approach consists of backing ideas with data that offer directional clues to react more effectively. Thanks to ten years of experience in developing AI technologies, we have been able to provide our scientists with the ability to democratize idea generation and get specific answers to specific problems.
Diamond’s ‘brain’ is equipped with a deep learning neural network technology, able to classify information from Kiromic’s Digital Library. This neural network technology is an extensive resource integrating clinical studies, genomic, and proteomic datasets. Diamond orchestrates all the raw data and produces new datasets for cancer target screening.
Diamond can also identify new genes (biomarkers, mutations, isoforms, neoepitopes, gene methylation status) highly and specifically expressed in targeted diseases. The software exploits its ‘intelligence’ to highlight these new genes across the entire patient population, mapping out the exact portion of the gene to elicit an immune response.
Additionally, Diamond can perform meta-analysis and convolution studies, integrating big data from experimental platforms, allowing an intuitive visualization of consistent and accurate results in a user-friendly fashion.
CancerDiff is differentiated by its user-friendly interface. The interface is a web-browser integrated suite, and once the user logs into the software, she/he can choose a starting analysis at ‘Tumor level,’ selecting the ‘Tumor Search’ check-box, and a popup menu with 38 cancers is provided.
We believe that Diamond is more accurate than other tools due to the implementation of robust statistics that compare healthy and pathologic gene expression. Other tools that utilize the TCGA database only provide information on healthy tissues. Moreover, Diamond offers a series of algorithms that single out the identified epitopes with a strong affinity for a given HLA haplotype, predicting how much those epitopes have an affinity for B and T cells.
Diamond is a powerful software that can accomplish several requests in the context of neoantigen discovery. The statistics implemented for the antigens identified are based on a weighted t-test conducted to compare the distributions of cancerous and healthy tissue expression data. The method is robust and effective.
We can do all of this through our integrated software: Diamond, and CancerSplice, solutions that are created ‘around’ scientists that require the ability to discover and explore new hypotheses based on supporting evidence in real-time. These powerful algorithms can quickly check an idea’s validity, with deeply applied methodological rigor, in real-time.
Our integrated software solution is how we have chosen to support critical business decisions concerning a therapeutic product, contributing to opening the possibility of better healthcare in the future.”
CSIO of Kiromic, Mr. Gianluca Rotino, stated:
“With the results displayed in these three posters, Kiromic has demonstrated the ability to position itself among the leaders in the development and application of Artificial Intelligence in immunology and immunotherapy in general.
The competitive advantage is multiple and disruptive.
Just think of the identification and development time of a biomarker with or without artificial intelligence.
Identifying specific isoforms of a gene for at least 38 cancers, using a manual strategy, could be equivalent (but not wholly equivalent) to finding two identical telephone numbers, comparing 38 different phone books from 38 different US cities.
We should take into consideration the possibility of a permutation system that encompasses all the combinatorial possibilities to find two similar chunks of genes, with different lengths, found in all of these 38 books.
This type of finding would be equivalent to a human working for 7.6690286748 × 1026195 weeks. And just for a single biomarker.
Thanks to the application of our AI, these times and related costs are reduced to a few weeks of computing with a reduced need for full time employees.
However, we believe that our AI platform will not just give us a significant advantage in the discovery process, but we believe it will also power our clinical trials.”
CFO of Kiromic, Mr. Tony Tontat, stated:
“Our drug discovery process has always had AI at its core.
Our investment and then our deployment of AI has allowed us to shorten the time required to reach our IND filing and, soon, our expected first-in-human clinical trial.
According to Reports.com market investments in AI in the Healthcare Market is expected to reach $8 Billion by 2022 with a CAGR of 52.68% (2017-2022).
We look forward to extending the lead in the market by continuing to invest in our AI technologies.”
----------------------------------
About Kiromic Artificial Intelligence
Kiromic Artificial Intelligence System, is a modular multilayer system composed of two main modules:
CancerDiff and Diamond AI.
CancerDiff is a computational tool designed to identify the upregulation of splice variants of specific target genes at the RNA level. The software can identify new effective and specific tumor targets for immunotherapies which exist as tumor selective isoforms that have null or very low presence in healthy tissues. The identification of these isoforms allows the researcher to specifically address unique biomarker variants expressed on the surface of cancer cells. Thus, CancerDiff can improve biomarkers screening selection and gene-target therapies.
Diamond is a robust suite of algorithms platform which is able to mine public and proprietary genomic and proteomic databases and to use machine learning to identify targets for immunotherapy. Diamond can map the most immunogenic fragments of a protein and can predict the immunogenicity of peptides and to calculate their affinities to different HLA types. Specifically, Diamond can predict the processing of peptides, their HLA binding, and their ability to activate T and B cell response. Other available prediction platforms can rarely perform all of these with accuracy. For example, not all peptides that are presented by target cells can efficiently trigger immune response. Our system can identify peptides that can do both.
----------------------------------------
About Kiromic
Kiromic Biopharma, Inc. (Nasdaq: KRBP) is a target discovery and gene-editing company utilizing a state-of-the-art artificial intelligence (AI) platform focused on unleashing the power of the patient’s own immune system to fight cancer.
Kiromic’s pipeline development is leveraged through the Company’s proprietary target discovery Artificial Intelligence engine called “DIAMOND.” Kiromic’s DIAMOND is big data science meeting target identification, dramatically compressing the man-years and the millions of drug development dollars needed to develop a live drug.
The Company maintains its HQ offices in the world’s largest medical center in Houston, Texas adjacent to the MD Anderson Cancer Center and the Baylor College of Medicine where Kiromic has ongoing collaboration with these Institutions.
Kiromic’s scientific achievements related to this endeavor can be seen in the six poster nominations awarded by the prestigious AACR at its annual meeting in April, 2021 (https://ir.kiromic.com/news-releases/news-release-details/kiromic-announces-6-posters-presented-american-association).
For more information, please visit the company’s website at www.kiromic.com.
Forward-Looking Statements
This press release contains “forward-looking statements” that are subject to substantial risks and uncertainties. All statements, other than statements of historical fact, contained in this press release are forward-looking statements. Forward-looking statements contained in this press release may be identified by the use of words such as “anticipate,” “believe,” “contemplate,” “could,” “estimate,” “expect,” “intend,” “seek,” “may,” “might,” “plan,” “potential,” “predict,” “project,” “target,” “aim,” “should,” "will” “would,” or the negative of these words or other similar expressions, although not all forward-looking statements contain these words. Forward-looking statements are based on our company’s current expectations and are subject to inherent uncertainties, risks and assumptions that are difficult to predict. Further, certain forward-looking statements are based on assumptions as to future events that may not prove to be accurate. These and other risks and uncertainties are described more fully in the section titled “Risk Factors” in the Company’s annual report on Form 10-K for the most recently completed fiscal year and subsequent reports filed after the date of the annual report with the Securities and Exchange Commission. Forward-looking statements contained in this announcement are made as of this date, and our company undertakes no duty to update such information except as required under applicable law.
