Personalised therapy in pancreatic carcinoma – a forthcoming opportunity?

Genetic tests

Genetic testing is a key element of precision medicine. It identifies mutations that may become therapeutic targets, as well as be a predictive factor to assess the prognosis and response to treatment. For PDAC, it remains even more important in view of the lack the predominant mutations or difficult-to-treat driver lesions (such as KRAS, TP53), and the tumour heterogeneity, which restrict the effectiveness of the commonly used therapy choices, making the disease often primary resistant to all forms of treatment [7, 8]. Therefore, the consecutive trails focusing on the identification of the predominant targetable driver lesions are ongoing, using whole genome sequencing.

In a study conducted by The Cancer Genome Atlas Research Network, 10% of 150 PDAC samples showed germline or somatic mutations in one of the DNA damage repair genes ATM, BRCA1, BRCA2, and PALB2, potentially sensitising PDAC to platinum-based chemotherapy or PARP inhibition. Low-prevalence alterations in several genes that could potentially be amenable to other targeted therapies were observed, including mutations in BRAF, PIK3CA, RNF43, STK11, and JAK1, as well as focal high-level amplifications in ERBB2. A single hypermutated sample harboured 19,957 mutations, including a mutation in POLE. This cancer may have a higher neo-antigen load; thus, this patient may be a candidate for immunotherapy. Excluding common KRAS or CDKN2A variants, 42% (63/150) of patients within this cohort had tumours with at least one genomic alteration that could potentially confer eligibility for ongoing clinical trials, and 25% of patients (38/150) had tumours with two or more such events, suggesting the need for genotype-driven combination therapy studies [9].

Waddell et al. performed whole-genome deep sequencing of 100 PDAC cases. Using the structural variant analysis, they classified PDAC into four subtypes with potential clinical relevance. A significant proportion of tumours contained amplifications and copy-number gains of known oncogenes, but most occurred at a low individual incidence, suggesting considerable diversity in the mechanisms involved in PDAC progression. Mutations in KRAS, TP53, SMAD4, CDKN2A, and ARID1A were common, in addition to multiple mutations at low prevalence. The recurrent mutations identified in KDM6A further highlight the role of chromatin modification, and the broader role of aberrant WNT signalling is associated with the relatively frequent inactivation of tumour suppressor genes, including ROBO1, ROBO2, SLIT2, and RNF43. Several of them also carried mutations in known therapeutic targets with available inhibitors (ERBB2, MET, FGFR1). Others included GATA6, PIK3CA, PIK3R3, and CDK6. The analysis also elucidated the co-segregation of impaired DNA repair genes (such as BRCA1, BRCA2, or PALB2) and genomic instability along with an association between favourable platinum-based treatment outcome. However, it remains questionable whether mutations in BRCA pathway components, both germline and somatic, as well as putative surrogate measures of deficiencies in DNA maintenance, including unstable genomes and the BRCA mutational signature, could be predictive of therapeutic response in the absence of BRCA or PALB2 mutations and help to identify potential responders to platinum and PARP inhibitor therapy, considering the fact that the current patient inclusion for clinical trials of PARP inhibitors, thought to target similar mechanisms, are mostly based on germline deleterious mutations of BRCA1 and BRCA2 [10].

Genomic instability in neoplastic cells is also an important determinant of disease outcome, given the fact that it has been shown to be a predictor for immune checkpoint inhibitor therapy. However, a study of 35 long-term pancreatic cancer survivors found that defective mismatch repair (MMR) DNA is uncommon and does not account for the survival benefit in patients with sporadic cancer [11]. Previous studies also suggest that defective MMR DNA is uncommon in sporadic pancreatic carcinoma. Fraune et al. confirmed that MSI occurs early in a small subset of PDAC [12].

Campbell et al. observed a process of continuous genetic rearrangements during the development and progression of pancreatic cancer due to genomic instability. For instance, metastasis samples of different sites from the same patient showed diverse genetic alternations at various loci, suggesting that an unstable genome confers additional characteristics to metastatic lesions derived from the same parental clone. This genetic diversity explains the heterogeneous response to treatment and the subsequent progression of some lesions while maintaining disease control in others after an initial therapy response [13].

The Know Your Tumor Registry Trial focused on biopsy-confirmed pancreatic cancers, including ductal adenocarcinoma, adenosquamous carcinoma, and acinar cell carcinoma. A retrospective analysis of 1082 patients showed that 26% patients had actionable molecular alterations that were not present in the other cancer subtypes. The group was divided into 2 subgroups: 46 patients received molecularly tailored therapy including pembrolizumab for MSI-H, olaparib for BRCA mutations, dabrafenib + trametinib for BRAF mutation, and larotrectinib or entrectinib for NTRK1 fusion; the second group consisted of 143 patients who received standard therapy without molecular profiling, usually chemotherapy – gemcitabine + nab-paclitaxel, FOLFIRINOX, or FOLFOX. The first group had improved overall survival (2.58 vs. 1.51 years, p = 0.0004) compared with the unmatched therapy group [14].

Patient-derived organoids

Several researchers have explored the use of patient-derived organoids (PDOs) in precision medicine for the treatment of pancreatic cancer. PDOs are miniature 3D cell cultures of tumour tissues cultivated in the laboratory from primary tumours obtained from patients. They are produced by cultivating adult stem cells with specific spatial arrangements in vitro. Currently, there is no universally standardised procedure for constructing organoids, and differences arise in the culture methods of various solid tumour organoids, particularly regarding the culture medium with some growth factors and inhibitor [15]. In recent years, they have recently emerged as a powerful model for personalised medicine applications. The number of organoids used for xenograft generation and molecular profiling is growing, and patient-derived organoid models are widely used as in vitro screening platforms with the potential to predict the best treatment options for individual patients. Retrospective trials have evaluated myriad chemotherapeutic agents introduced into therapy only after the initiation of organoid culture. Many studies have been conducted on drug screening using PDOs; however, it remains unclear whether and how this approach could be tailored to patients’ clinical responses and, more importantly, whether these results can be translated into clinical decision-making strategies [16].

Huang et al. developed three-dimensional culture conditions to induce differentiation of human pluripotent stem cells into exocrine progenitor organoids that form ductal and acinar structures in culture and in vivo. The study confirmed that expression of mutant KRAS or TP53 in progenitor organoids induced mutation-specific phenotypes in culture and in vivo. In particular, expression of TP53^R175H induced cytosolic localisation of SOX9, which was associated with lethality. In addition, culture conditions were established for clonal generation of tumour organoids from freshly resected PDAC. Tumour organoids retained the differentiation status, histoarchitecture, and phenotypic heterogeneity of the primary tumour and retained patient-specific physiological changes, including hypoxia, oxygen consumption, epigenetic marks, and differential sensitivity to EZH2 inhibition. Thus, PDOs have become an option for drug screening before selecting appropriate therapy [17].

Tiriac et al. described a library of 66 PDO cultures obtained from fine-needle biopsy, surgical resection, and postmortem PDAC samples collected at multiple clinical institutions. Using molecular characterisation of the PDO genome and transcriptome, they identified hallmarks of PDAC, including the oncogenic alleles ERBB2^S310F, MAP2K1^Q58-E62del, and PIK3CA^E110del. In addition, high concordance was found between primary tumour samples and paired PDOs when sufficient tumour cell density was observed in the patient sample. By establishing a PDAC-specific drug testing pipeline with PDOs, they demonstrated that drug sensitivity profiles for commonly used chemotherapeutic agents, such as gemcitabine, nab-paclitaxel, irinotecan (SN-38), 5-fluorouracil (5-FU), and oxaliplatin, could be generated for each PDO within a time frame that allows clinical application. Furthermore, in a retrospective analysis of patients with advanced PDAC from whom PDOs were generated, the PDOs chemotherapy sensitivity profile reflected the patient’s response to therapy administered according to the PDO trial results [18].

Driehuis et al. used 30 patient-derived organoid lines (PDOs) from tumours arising in the pancreas and distal bile duct. PDOs recapitulated tumour histology and contained genetic alterations characteristic for pancreatic cancer. In vitro testing of a panel of 76 therapeutic agents, including chemotherapeutics currently used in the treatment of PDAC, or multiple agents targeting microtubules, aurora kinase A (AURKA), phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA), and topoisomerase-1 (TOP1), revealed sensitivities currently not exploited in the clinic. Also, for 4 patients, the available clinical data allowed for a comparison of the response to treatment of both the patient and the matching PDO, and the results were consistent. Finally, they used pancreatic cancer organoids to evaluate the potential of a new targeted agent. They confirmed that PRMT5 inhibition (EZP015556) effectively targets MTAP-negative tumours, a common alternation in 15% of all human cancers, including pancreatic cancer [19].

In the study by Sharick et al., noninvasive optical metabolic imaging (OMI) of cellular heterogeneity was characterised in the context of breast and pancreatic cancer PDOs. OMI of PDAC organoids, previously studied by Walsh et al., is considered an attractive way to study the response to new drugs [20]. Organoids were treated with the same drugs as the patient’s prescribed regimen (usually consisting of a combination of gemcitabine, 5-FU, oxaliplatin, nab-paclitaxel, SN-38, docetaxel, cyclophosphamide, doxorubicin, trastuzumab, or pertuzumab), and OMI heterogeneity measurements were compared with patient outcomes. The OMI study identified subpopulations of cells with diverse and dynamic responses to treatment in living organoids without the use of markers or dyes. The OMI results of organoids were consistent with the long-term therapeutic response in patients [21].

Seppälä et al. obtained PDOs from tissues secured during surgical resection or biopsy from patients previously receiving cytotoxic chemotherapy (8 cycles of FOLFIRINOX regimen). They demonstrated the capacity of rapidly established PDO disease models to serve as reliable, patient-specific prognostic biomarkers of clinical chemotherapeutic response, and their generation was sufficiently short to be used in clinical practice to guide therapy selection. Additionally, using next-generation sequencing, they detected unique, clinically relevant genetic mutations after surgical resection that complemented the selection of appropriate therapy [22].

mRNA vaccines

Another option for personalising the therapeutic approach in pancreatic malignancies is the use of mRNA vaccines. PDAC seems to be almost completely insensitive (< 5% response rate) to immune checkpoint inhibitors, which is partially attributed to a low tumour mutation rate while generating few neoantigens, mutation-generated proteins absent in healthy tissues that mark cancers as foreign to T cells. Therefore, PDAC tissue remains weakly antigenic with few infiltrating T cells [8, 23]. Based on observations, it has been established that in people with long-term survival after PDAC detection, there is an increase in spontaneous T-cell responses to neoantigens distinct for the tumour in a specific patient [24, 25]. Thus, research has begun to investigate whether a personalised vaccine can stimulate neoantigen-specific T cells and provide therapeutic benefits in the adjuvant setting. In view of the fact that the mRNA vaccine technology facilitates rapid delivery of personalised neoantigen vaccines, which can be fully integrated into the standard clinical management in oncology, eliminating micrometastases and delaying disease relapse.

Rojas et al. showed that adjuvanted autogen cevumeran, an individualised neoantigen vaccine based on uridine mRNA–lipoplex nanoparticles, in combination with atezolizumab and mFOLFIRINOX, is safe, feasible, and generates significant neoantigen-specific T cells in 50% of unselected patients with resectable PDAC. T cells expanded after vaccination persisted for up to 2 years despite post-vaccination treatment with an mFOLFIRINOX regimen [26]. A phase II trial is currently underway (NCT05968326).

There are some other phase I studies evaluating the efficacy of mRNA vaccines in pancreas cancer treatment. One Shanghai trial assessed the anti-tumour activity of KRAS neoantigen mRNA vaccine (ABO2102) alone and in combination with toripalimab (anti-PD-1 monoclonal antibody) among participants with KRAS-mutated advanced pancreatic cancer (NCT06577532). Other Chinese trials concern a neoantigen personalised mRNA tumour vaccine combined with adebrelimab (a PD-L1 humanised monoclonal antibody) (NCT06156267), camrelizumab (an anti-PD-1 antibody) (NCT06326736), ipilimumab with gemcitabine + capecitabine (NCT06353646), or undefined PD-1 inhibitor (NCT06496373) in patients with surgically resected pancreatic adenocarcinoma. The results of a completed phase 1 study of mRNA-5671/V941 as a monotherapy and in combination with pembrolizumab in participants with KRAS mutant advanced or metastatic non-small cell lung cancer, colorectal cancer, or pancreatic adenocarcinoma are not currently published (NCT03948763).

An interesting proposal was suggested by Masum et al., presenting the results of a study in which they constructed an mRNA vaccine using an in silico approach – they analysed S100 family proteins, all of which are major activators of advanced glycation end product receptors. They used immunoinformatic tools, including analysis of physicochemical properties, structural prediction and validation, molecular docking study, in silico cloning, and immunological simulations. In the structural evaluation, the designed vaccine proved to be stable and characterised by high affinity for TLR-2 and TLR-4 receptors. In immunological simulations, the appearance of memory B cells and T cells was observed, with increased levels of helper T cells and immunoglobulins (IgM and IgG). The results of the study are promising, but clinical trials are necessary to establish its therapeutic efficacy [27].