827 Streamlining design of safe and effective TCR therapies with AI

Background

Adoptive cell therapies with T lymphocytes expressing engineered T cell receptors (TCRs) are one of the most promising approaches to cancer therapy.1 However, the experimentally driven development of novel TCR therapies is limited by the enormous biological variability of peptide:Human Leukocyte Antigen:TCR (pHLA:TCR) complexes. The in silico methods hold the promise to streamline the discovery of novel TCR therapies by reducing costs and time of laboratory research. In particular, the prediction of TCR binding to a target antigen, as well as the prediction of TCR off-target toxicity2 can provide useful insights supporting the development of safe therapies. We aimed at the development of an experimentally validated AI model of pHLA:TCR binding that will help to prioritize and reduce the number of in vitro assays necessary to discover novel TCRs for cancer therapies.

Methods

The limiting factor of successful pHLA:TCR binding modeling is data availability and completeness of TCR characterization. To address this issue, we are building an oncological pHLA:TCR database with paired alpha and beta chain TCR sequences. We are collecting and sequencing tumor and normal samples from 100 cancer patients, as part of an observational clinical trial. Those data are then screened with the Ardigen's ArdImmune Vax platform3 4 to select immunogenic epitopes. T cells that bind those epitopes are subsequently sorted and used to generate TCR sequencing data at single-cell resolution. We use data-driven and simulation-based models to extract insights about the dynamics of a pHLA:TCR system to predict the binding probability and explain the inference made by the model.

Results

We optimized our data collection pipeline for the cost-efficient acquisition of a large oncological pHLA:TCR dataset. These data will enable us to build efficient models to streamline the development of TCR therapies against cancer.We benchmarked our modeling approach for pHLA:TCR binding against existing solutions5–7 on publicly available data. We also show how focus on model explainability facilitates the detection of model inconsistency of uncertain predictions by expert inspection. Our toxicity assessment solution2 extends the applicability of our system to the prediction of TCR safety profile.

Conclusions

The presented work shows perspectives and limitations of AI-aided TCR therapy development. We present results for our pHLA:TCR binding model, a TCR-toxicity-screening solution, and the study design of our observational clinical trial. Our growing database of pHLA:TCR interactions will enable us to develop highly predictive pHLA:TCR binding models, in particular for oncological targets.

Acknowledgements

We acknowledge funding through the project “Creating an innovative AI-based (Artificial Intelligence) IN SILICO TECHNOLOGY TCRact to launch a NEW SERVICE for designing and optimizing T-cell receptors (TCR) for use in cancer immunotherapies” cofunded by European Regional Development Fund (ERDF) as part of Smart Growth Operational Programme 2014–2020.

References

Farkona S, Diamandis EP, Blasutig IM. Cancer immunotherapy: the beginning of the end of cancer? BMC Med 2016;14:73. PMCID: PMC4858828.

Murcia Pienkowski VA, Mazzocco G, Niemiec I, Sanecka-Duin A, Krol P, Myronov O, Skoczylas P, Kaczmarczyk J, Blum A. Off-target toxicity prediction in cellular cancer immunotherapies [Internet]. Cytotherapy. 2021;S96. Available from: http://dx.doi.org/10.1016/s1465324921004229.

Stepniak P, Mazzocco G, Myronov A, Niemiec I, Gruba K, Skoczylas P, Sanecka-Duin A, Drwal M, Kaczmarczyk J. AI-augmented design of effective therapeutic cancer vaccines and adoptive cell therapies. Journal For Immunotherapy Of Cancer. Bmc Campus, 4 Crinan St, London N1 9xw, England; 2019.

Mazzocco G, Niemiec I, Myronov A, Skoczylas P, Kaczmarczyk J, Sanecka-Duin A, Gruba K, Król P, Drwal M, Szczepanik M, Pyrc K, Stȩpniak P. AI aided design of epitope-based vaccine for the induction of cellular immune responses against SARS-CoV-2. Front Genet. 2021;12:602196. PMCID: PMC8027494.

Weber A, Born J, Rodriguez Martínez M. TITAN: T-cell receptor specificity prediction with bimodal attention networks. Bioinformatics. 2021;37(Suppl_1):i237–i244. PMCID: PMC8275323.

Springer I, Besser H, Tickotsky-Moskovitz N, Dvorkin S, Louzoun Y. Prediction of specific TCR-peptide binding from large dictionaries of TCR-Peptide Pairs. Front Immunol 2020;11:1803. PMCID: PMC7477042.

Jurtz VI, Jessen LE, Bentzen AK, Jespersen MC, Mahajan S, Vita R, Jensen KK, Marcatili P, Hadrup SR, Peters B, Nielsen M. NetTCR: sequence-based prediction of TCR binding to peptide-MHC complexes using convolutional neural networks [Internet]. Available from: http://dx.doi.org/10.1101/433706

AI Aided Design of Epitope-Based Vaccine for the Induction of Cellular Immune Responses Against SARS-CoV-2

The heavy burden imposed by the COVID-19 pandemic on our society triggered the race toward the development of therapies or preventive strategies. Among these, antibodies and vaccines are particularly attractive because of their high specificity, low probability of drug-drug interaction, and potentially long-standing protective effects. While the threat at hand justifies the pace of research, the implementation of therapeutic strategies cannot be exempted from safety considerations. There are several potential adverse events reported after the vaccination or antibody therapy, but two are of utmost importance: antibody-dependent enhancement (ADE) and cytokine storm syndrome (CSS). On the other hand, the depletion or exhaustion of T-cells has been reported to be associated with worse prognosis in COVID-19 patients. This observation suggests a potential role of vaccines eliciting cellular immunity, which might simultaneously limit the risk of ADE and CSS. Such risk was proposed to be associated with FcR-induced activation of proinflammatory macrophages (M1) by Fu et al. (2020) and Iwasaki and Yang (2020). All aspects of the newly developed vaccine (including the route of administration, delivery system, and adjuvant selection) may affect its effectiveness and safety. In this work we use a novel in silico approach (based on AI and bioinformatics methods) developed to support the design of epitope-based vaccines. We evaluated the capabilities of our method for predicting the immunogenicity of epitopes. Next, the results of our approach were compared with other vaccine-design strategies reported in the literature. The risk of immuno-toxicity was also assessed. The analysis of epitope conservation among other Coronaviridae was carried out in order to facilitate the selection of peptides shared across different SARS-CoV-2 strains and which might be conserved in emerging zootic coronavirus strains. Finally, the potential applicability of the selected epitopes for the development of a vaccine eliciting cellular immunity for COVID-19 was discussed, highlighting the benefits and challenges of such an approach.

Abstract 6541: Accounting for immune escape mechanisms in personalized and shared neoantigen cancer vaccine design

The purpose of this study is to evaluate the importance of immune escape mechanisms (IEMs) when designing cancer immunotherapeutic strategies.

Cancer IEMs induced by negative selective pressure on tumor cells can have dramatic effects on immunotherapy treatment outcome. IEMs occurring during cancer evolution have been shown to facilitate loss of mutation-associated neoantigens (Anagnostou et al. Cancer Discov 2017). These mechanisms include: Loss of Heterozygosity (LOH), underexpression of HLA genes and mutations in the neoantigen presentation machinery.

During cancer evolution, the initially arising clonal neoantigens may be shared across many patients, followed by predominantly subclonal private neoantigens. The propensity of shared neoantigens to be clonal makes them promising candidates for the development of off-the-shelf cancer vaccines.

In order to optimize patient treatment strategies and qualify patients for either shared or private neoantigen-directed therapy, it is crucial to correctly assess the indication-specific IEMs and neoantigen landscape in both hot and cold tumors.

We carry out an extensive analysis of tumor neoantigen landscape from different indications based on TCGA, on Parkhurst et al. Cancer Discov. 2019 and other available data for selected indications (including MEL, NSCLC, CRC, and GBM). Using bioinformatics methods we investigate IEMs and their potential effects on the plasticity of neoantigen repertoires including enrichment of immunogenic neoantigens within the predicted vaccine composition, computed per patient. Immunogenicity is assessed using advanced deep learning models for neoantigen presentation and functional T-cell response prediction (Mazzocco et al., poster presented at SITC 2019).

We present the results of the comparison between different cancer indications including CRC, MEL, NSCLC and GBM using TCGA patient data and GI-tract cancer patients with validated T-cell responses. We present the frequency of particular immune evasion events, pointing out the differences in the number and quality of immunogenic neoepitopes (both shared and private) under those conditions and provide summary statistics.

As a result, we draw conclusions for vaccine design strategies, which are informed by the investigation of immune evasion events and are tested on selected datasets including experimentally-validated neoantigen responses.

Citation Format: Alexander Myronov, Iga Niemiec, Katarzyna Gruba, Giovanni Mazzocco, Anna Sanecka-Duin, Piotr Skoczylas, Michał Drwal, Jan Kaczmarczyk, Piotr Stepniak. Accounting for immune escape mechanisms in personalized and shared neoantigen cancer vaccine design [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 6541.

Elective lung resection increases spatial QRS-T angle and QTc interval

BACKGROUND

Lung resection changes intra-thoracic anatomy, which may affect electrocardiographic results. While postoperative cardiac arrhythmias have been recognized after lung resection, no study has documented changes in vectorcardiographic variables in patients undergoing this surgery. The purpose of this study was to analyse changes in spatial QRS-T angle (spQRS-T) and corrected QT interval (QTc) after lung resection.

METHODS

Adult patients undergoing elective lung resection under general anaesthesia were studied. The patients were allocated into four groups: those undergoing (1) left lobectomy (LL); (2) left pneumonectomy (LP); (3) right lobectomy (RL); and (4) right pneumonectomy (RP). The spQRS-T angle and QTc interval were measured one day before surgery (baseline) and 24, 48 and 72 h after surgery.

RESULTS

Seventy-one adult patients (47 men and 24 women) aged 47-80 (65 ± 7) years were studied. In the study group as a whole, lung resection was associated with significant increases in spQRS-T (p < 0.001) and QTc (p < 0.05 at 24 and 48 h and p < 0.01 at 72 h). The greatest changes were noted in patients undergoing LP. Postoperative atrial fibrillation (AF) was noted in 6.4% of patients studied, in whom the widest spQRS-T angle and the most prolonged QTc intervals were also noted.

CONCLUSIONS

Lung resection widens the spQRS-T angle and prolongs the QTc interval, especially in patients undergoing LP. While postoperative AF was a relatively rare complication after lung resection in this study, it was associated with the widest spQRS-T angles and most prolonged QTc intervals.

The presence and clinical implications of α-2,6-galactose-linked sialic acids in non-small-cell lung cancer brain metastases--preliminary study

Brain metastases (BM) in non-small-cell lung cancer (NSCLC) patients present an increasing clinical challenge. Identifying biomarkers which specifically identify patients at high risk of BM may improve their early diagnosis, which is crucial for surgical and radiotherapeutic treatment outcome. Alpha-2,6-sialyltransferase (α-2,6-ST) and the primary product of its activity, alpha-2,6-galactose-linked sialic acids (α-2,6-GalSA) have been found responsible for the adhesion of tumor cells to the brain vessels' endothelium and enabling their transmigration through the blood-brain barrier in brain metastatic tumors. The aim of the study was to investigate by histochemical method the presence and possible role of α-2,6-GalSA in the formation of brain metastasis in NSCLC. In the screening phase 76 metastatic brain tumors were stained for α-2,6-GalSA and the second phase involved an identical staining of 20 primary tumors of patients who had their primary tumors treated with surgery or definite radiochemotherapy yet who later developed BM. The results were compared to a control group of 22 patients treated with surgery for NSCLC and who survived 5 years without the recurrence of disease. Alpha-2,6-GalSA presence was found to be down-regulated in poorly differentiated tumor types, whereas majority of differentiated tumors overexpressed it. This was statistically significant for both BM and the primary tumors. The expression of α-2,6-GalSA remained stable in primary and metastatic tumor pairs, however, no statistically significant differences were observed between study and control groups. Within the study group, a higher α-2,6-GalSA expression was associated with better overall survival, but not all statistical models found this result significant. Further studies are recommended to validate these findings.

Urban vs. rural patients. Differences in stage and overall survival among patients treated surgically for lung cancer

INTRODUCTION

Besides the undoubted influence of risk factors on morbidity and survival time, there are also other environmental factors, such as awareness of the prevalence of risk factors and the availability of modern diagnosis and treatment methods.

OBJECTIVE

To evaluate differences in lung cancer 5-year overall survival rates between urban and rural patients hospitalized in the Department of Thoracic Surgery of the Medical University in Lublin, Poland, and possible influence of several risk factors on these rates.

MATERIALS AND METHODS

The analysis was based on 125 lung cancer patients who underwent surgical procedures in years 2006-2007 and who agreed to take part in the survey. The study aimed at recognition of the health situation and selected demographic traits of people who had been treated surgically for lung cancer. The differences were evaluated between rural and urban inhabitants in gender, age, lung function, smoking habits, exposure to risk factors at work, family history of cancer, staging of the disease, histological type of cancer, post-surgical treatment, and their possible influence on overall survival.

RESULTS

The results showed that the only noted differences between urban and rural population were in tobacco smoking and lung function. Survival rates were very similar and did not differ from the European average.

CONCLUSIONS

The assumption that Polish rural patients are presenting with later cancer stages at the time of diagnosis, and have worse chances for survival, has become invalid in modern times.

[Active and passive smoking among students of Polish high schools]

The survey was conducted among 29540 students in different secondary schools. All interviewers were 17.42 +/- 1.14 years old. The aim od this work was to define the relation between smoking adolescent and their health. The group of smokers consisted of 5112 (60.26%) men and 3370 (33.70%) women. The main causes of initiating smoking reported by respondents included curiosity--(47.91%) followed by stress (39.82%) and alcohol (6.77%). Smoking tobacco by parents at home results in inspiring by their children the unwholesome ingredients of tobacco smoke. Frequent staying among tobacco smokers significantly influences decision of starting or giving up smoking tobacco (59.24%). The obtain results may be helpful in preparing effective educational and preventive project among students, which can eliminate or lower the influence of factors predisposing to start and continue smoking tobacco. Special programs should be worked out for children showing early health consequences of smoking cigarettes. Giving up smoking is possible for "young" smokers, if they have not developed the symptoms of addiction.

[Smoking among secondary school students in Stalowa Wola]

Smoking is recognized as major, the most prevalent, and at the same time the only removable factor of many diseases. The major causes of starting smoking are: stress, curiosity and staying among people who smoke tobacco. The research was conducted among students of secondary school in Stalowa Wola in the form of an anonymous poll. The sample group included 2763 people. The group of smokers included 334 people (12.09%). Results of the research were subjected to a precise statistical analysis. The fact that 24.55% young age smokers smoke more than 20 cigarettes a day is worth stressing. In spite of short time of tobacco smoking, the investigated group has reported some negative health ailments. An attempt was made to identify the combination of features affecting smoking, including sex, origin, company of smokers, economic conditions (p < 0.05). The statistical analysis revealed a significant correlation between the number of cigarette smoked, economic conditions and the intention to stop smoking (p < 0.05). The obtained results may be helpful in preparing effective educational and preventive project among students, which can eliminate or lower the influence of factors predisposing to start and continue smoking tobacco.