KRAS-driven Tumorigenesis and KRAS-driven Therapy in Pancreatic Adenocarcinoma

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is associated with significant morbidity and mortality and is projected to be the second leading cause of cancer-related deaths by 2030. Mutations in KRAS are found in the vast majority of PDAC cases and plays an important role in the development of the disease. KRAS drives tumor cell proliferation and survival through activating the MAPK pathway to drive cell cycle progression and to lead to MYC-driven cellular programs. Moreover, activated KRAS promotes a pro-tumorigenic microenvironment through forming a desmoplastic stroma and by impairing anti-tumor immunity. Secretion of GM-CSF and recruitment of myeloid-derived suppressor cells and pro-tumorigenic macrophages results in an immunosuppressive environment while secretion of SHH and TGF-beta drive fibroblastic features characteristic of PDAC. Recent development of several small molecules to directly target KRAS mark an important milestone in precision medicine. Many molecules show promise in preclinical models of PDAC and in early phase clinical trials. In this review, we discuss the underlying cell intrinsic and extrinsic roles of KRAS in PDAC tumorigenesis, the pharmacologic development of KRAS inhibition, and therapeutic strategies to target KRAS in PDAC.

Introduction:

The advance of targeted- and immuno-therapy has changed the therapeutic landscape across clinical oncology. However, these interventions have had limited clinical benefit in pancreatic ductal adenocarcinoma (PDAC), which portends a poor prognosis with a 5-year overall survival rate of 12% [1]. Advanced disease at the time of diagnosis and resistance to cytotoxic chemotherapies contribute to poor survival. Additionally, PDAC infrequently possesses druggable genomic alterations, such as mutations in BRCA, PALB2, or fusions with NTRK or NRG1. KRAS mutations, however, are implicated in over 90% of PDACs. Mutations in KRAS drive tumorigenesis while additionally altering the local inflammatory milieu. Though previously regarded as undruggable, the recent clinical breakthrough of active KRAS^G12C inhibitors has provided new promise against one of the most prevalent oncogenes. Although KRAS^G12C mutations are uncommon in PDAC, the new ability to target this oncogene is an enormous scientific feat that opens the door to further develop KRAS inhibitors relevant to PDAC. In this review, we discuss the impact of KRAS activation on tumorigenesis and the preclinical and translational efforts to target KRAS in PDAC.

KRAS activation in PDAC tumorigenesis

KRAS (Kirsten rat sarcoma viral oncogene homolog) is a small GTPase that cycles between a GDP-bound “off”-state and a GTP-bound “on”-state through interactions with GTPase-activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs), respectively. Activation of receptor tyrosine kinases (RTKs) promotes recruitment of Src homology, Collagen family adaptor protein (SHC) and growth factor receptor-bound protein 2 (GRB-2). This scaffold recruits the GEF, son of sevenless-1 (SOS1) to help activate KRAS [2]. To serve as a conduit for extracellular signals via receptor tyrosine kinases, KRAS needs to be anchored to the cellular membrane. This is achieved by farnesylation and carboxymethylation of the C185 residue to allow insertion into the lipid bilayer (Figure 1A). Moreover, the presence of a polybasic stretch of amino acids in the highly variable region preceding C185 allows for interaction with both acidic phospholipids and prenylation-binding proteins to ensure proper localization of KRAS to the plasma membrane [3–5]. Mutations frequently impact the G12, G13, and Q61 codons of KRAS, and 90% of mutations occur at G12 in PDAC [6]. These mutations localize to the catalytic GTPase site of KRAS, impairing GTP hydrolysis [7, 8]. Notably, oncogenic KRAS is not permanently locked in an “on”-state but, rather, is in a prolonged activated state that is influenced by growth factors and other modifiers, albeit to differing magnitudes [9, 10]. Moreover, different mutations within KRAS differentially impacts the structure of KRAS and influence its interactions with GAPs [11]. For an in-depth discussion on the impact of mutations and nucleotide binding on the structure of KRAS, the authors direct interested readers to the review by Tatu Pantsar [12].

Figure 1.

KRAS signaling and KRAS-directed therapy

A. Schematic outline of KRAS gene-structure, highlighting hotspot mutations and farnesylated cysteine. B. Protein-interaction network evaluating high-confidence interactions with KRAS based on experimentally determined, curated, and co-expression databases. C. Outline of cell-intrinsic factors governed by KRAS signaling. Therapies undergoing clinical evaluation are denoted. D. Outline of cell-extrinsic factors governed by KRAS signaling. HVR – high variable region. CAFs – cancer-associated fibroblasts. MDSC – Myeloid-derived suppressor cells. TCR – T-cell receptor. GAP – GTPase activating protein. Created with BioRender.com

Upon GTP-binding, the switch I and switch II domains of KRAS undergo conformational change to interact with downstream effectors [13]. Effector proteins are defined as proteins that contain a RAS-binding or RAS-association domain. Classical effector proteins include RAF, PI3K, and RALGDS, but there are many others (Figure 1B). Bioinformatic evaluation predicts over 50 different effectors and biochemical purification or proximity labeling implicates hundreds of effector-complexes [14–16]. These effectors potentiate downstream signaling cascades to alter cellular proliferation, survival, metabolism, and motility that ultimate drive pancreatic tumorigenesis [17, 18].

Of the multiple downstream effectors of RAS, expression of oncogenic BRAF to activate the RAF-MEK-ERK pathway in mice is sufficient to drive the development of PDAC, while expression of oncogenic PIK3CA to activate the PI3K-AKT-mTOR pathway results in a grossly normal mouse pancreas [19]. These findings highlight the importance of the RAF-MEK-ERK pathway, rather than the P13K-AKT-mTOR pathway, in the development of PDAC, a finding that is corroborated by the infrequent (~1%) mutational and copy number changes in the PI3K pathway found in human PDAC [6, 20].

Additionally, other mouse models of PDAC and KRAS have provided biologic insight into the development of PDAC. These in vivo models help to understand the underlying features of tumorigenesis given their similarities in their tumor microenvironment and ability to metastasize [21]. In the setting of PDAC, KRAS mutations are early oncogenic events that lead to the development of precancerous pancreatic intraepithelial neoplasia (“PanIN”) lesions [21–23]. Autopsy studies reveal that the prevalence of PanIN lesions may be as high as 70% in healthy persons, with lesions that arise as early as the second decade of life [24–26]. Molecular evaluation using laser micro-dissection of individual PanINs has identified the presence of KRAS mutations within these precursor lesions and recent studies suggest that hundreds of individual PanIN lesions can be present within a single person, all of which are likely derived from independent KRAS mutations [23, 27].

KRAS signaling results in phosphorylation and activation of ERK. ERK drives cellular growth and PDAC progression through increased MYC stability, ultimately enabling unregulated cellular proliferation. This is mechanistically accomplished by 1) activating MYC via phosphorylation and 2) stabilizing MYC by phosphorylation and deactivation of FBXW7, an E3-ligase that promotes proteasomal degradation of MYC [28, 29]. In preclinical models of PDAC, decreasing MYC stability via pharmacologic or molecular inhibition impairs tumor growth [30, 31]. Conversely, increases in MYC copy number is associated with metastatic disease in preclinical models [32]. Evaluation of metastatic PDAC lesions in patient samples correspondingly shows amplification in 30% of metastatic samples [33]. Although signaling flux impinges on ERK activation to help stabilize MYC, ERK provides negative feedback via phosphorylation of SOS1 to impair SOS1 interaction with EGFR [34, 35]. However, mutated KRAS is presumably less sensitive to this negative regulation, shifting the homeostatic signaling state to allow for continual signal propagation to enforce ongoing cellular growth.

KRAS and the PDAC Tumor Microenvironment

Mutated KRAS not only drives intrinsic cellular proliferation, but also acts to modulate the surrounding tumor microenvironment (TME) to help drive PDAC tumorigenesis. PDAC is characterized by an immunosuppressive milieu with a highly desmoplastic stroma [36]. These microenvironment changes occur as early as the PanIN stage. During the development of PanINs, pancreatic acinar cells harboring mutant-KRAS drive the expression of intercellular adhesion molecule-1 (ICAM-1) and granulocyte-macrophage colony-stimulating factor (GM-CSF) to recruit macrophages and myeloid-derived suppressor cells (MDSCs) to dysplastic lesions (Figure 1D). Extrinsic KRAS signaling leads to polarization of tumor-associated macrophages into the tumorigenic M2 phase, which drives tumorigenesis and correlates with poor prognosis [37, 38]. Together, these components function to impair anti-tumor immunity and promote immune evasion [39–41]. Infiltrating myeloid cells subsequently secrete TNF, promoting tumor cell secretion of chemokines such as CXCL1 and CXCL2 via the NF-kappa-B pathway [42]. These chemokines further potentiate the recruitment and reprogramming of MDSCs to exclude and dampen cytolytic CD8+ T-cell function, resulting in an immunologically cold tumor. Additionally, tumor cell production of CCL5 serves as a chemokine for homing suppressive CD4+ T-regulatory (Treg) cells to tumors [43]. Recently, an evaluation of human PDAC samples with single-cell mass cytometry corroborated this preclinical immunoregulatory circuit and demonstrated that this regulatory network is dependent on mutation of TP53 [44]. Disruption of this immunosuppressive circuit by impairing macrophage function (by neutralizing ICAM-1 antibodies or pharmacologic CSF1R inhibition), depleting CD4+ T-cells (by neutralizing antibodies), or compromising chemokine signaling (by genetic ablation of CXCL2, CXCL1 or CCL5 or pharmacologic inhibition of CCR2 or CCR5) results in reduced tumorigenesis and increased sensitization to chemotherapy in murine models, highlighting the crucial role of immune cells in initiating tumorigenesis and modulating therapeutic responses [41, 44–48]. Additionally, genetically removing KRAS in PDAC cell lines significantly impairs their ability to initiate tumors in immunocompetent mouse models more significantly than in immunocompromised mouse models, underlining the central importance of KRAS in immune evasion and tumor maintenance [49].

The dense desmoplastic stroma that is phenotypically intertwined with PDAC tumors is also driven by mutant-KRAS. Pancreatic stellate cells (PSCs) comprise 4% of all pancreatic cells but play an essential role in regulating the extracellular matrix (ECM) and fibrosis in the setting of inflammation. In their quiescent state, these cells maintain ECM homeostasis and are characterized as lipid rich cells that store vitamin A. When activated, PSCs lose their lipid and vitamin A contents and differentiate into myofibroblastic cancer-associated fibroblasts (myCAFs) or inflammatory cancer-associated fibroblasts (iCAFs). myCAFs express alpha-smooth muscle Actin (α-SMA) and remodel the tumor microenvironment while iCAFs secrete interleukin-6 (IL-6), leukemia inhibitory factor (LIF) and other cytokines to promote a local immunosuppressive environment [50–52]. myCAFs are typically found in the proximity of tumor cells, while iCAFs can be more scattered. KRAS-mutated tumor cells secrete sonic hedgehog (SHH), interleukin-1 (IL-1), and transforming growth factor-beta (TGF-β) to activate PSCs to drive subsequent desmoplasia and a fibroinflammatory stroma [22, 53, 54]. Additionally, antigen-presenting CAFs (apCAFs) were recently discovered through single-cell transcriptomic evaluation of murine and human PDAC. These cells were found to express MHCII without costimulatory molecules and are postulated to modulate CD4+ T-cell function, potentially through deactivating CD4+ T-cells [55]. Interrogation of myCAF function was explored through depleting SHH, which ultimately decreased desmoplasia and increased sensitivity to gemcitabine. However, tumors that lacked SHH or α-SMA+ myCAFs were characterized as more undifferentiated and aggressive, revealing the complex effects that desmoplasia can have on tumor biology and therapeutic intervention [56–58]. In addition to fibrosis and immunosuppression, activated PSCs and CAFs can reciprocate paracrine signaling to engage the incipient KRAS-mutated tumor cells. Secretion of IGF-1 and GAS6 from CAFs activates AKT-mediated signaling in tumor cells, while LIF from CAFs activate the STAT3 pathway [59, 60]. Ongoing research aimed at dissecting these cellular relationships in the context of CAF heterogeneity and plasticity may provide strategies to leverage for therapeutic benefit.

Altogether, KRAS drives both intrinsic and extrinsic features of tumorigenesis. This, in combination with the frequency of KRAS mutations, underscores why inhibiting KRAS will be transformative for clinical oncology, particularly for PDAC patients.

Direct KRAS inhibition in PDAC

The smooth surface of KRAS along with its high affinity for GTP/GDP has made targeting this oncogene very challenging, and indeed, RAS-family proteins were initially deemed to be “undruggable”. Given the initial challenges of direct RAS inhibition, indirect inhibitors provided an alternative approach to targeting the MAPK pathway. These initial methods included using farnesyltransferase inhibitors and MEK inhibitors. Conceptually, farnesyltransferase inhibitors were thought to be a viable strategy and proved to inhibit RAS localization. Unfortunately, the farnesyltransferase inhibitor, tipifarnib, showed no added clinical benefit in PDAC patients when compared to gemcitabine monotherapy [61]. One possible cause of the limited efficacy was thought to be from compensatory geranylgeranylation of KRAS in the setting of farnesyltransferase inhibition [62]. Subsequent evaluation with dual farnesyl/geranylgeranyl transferase inhibitors demonstrated promising preclinical activity [63]. Utilizing MEK inhibitors to impair downstream signaling was also thought to be a viable option. However, when trametinib was combined with chemotherapy in a randomized trial, no clinical benefit was observed compared to chemotherapy alone [64].

Targeting KRAS and MAPK in PDAC remained a challenge. However, in 2013, the Shokat lab discovered an allosteric binding pocket near the switch II region of KRAS [65]. This pocket was revealed through disulfide-tethering of the G12C residue of mutant-KRAS and led to the development of compounds that could covalently bind to the newly mutated cysteine residue while interacting with the switch II pocket to alter affinity for GTP/GDP [65]. This spurred investigations that identified improved molecules (ARS-853, ARS1620, MRTX849, AMG510, JDQ443) that could trap KRAS^G12C in the inactive state in vitro and in vivo [66–71]. ARS-853 was modified to create ARS1620 by addition of a quinazoline ring to interact with the Switch II pocket [68]. Doing so resulted in a 10-fold increase in potency (IC50 of 120 nM compared to 1,700 nM). AMG510 uses a similar quinazoline core as ARS1620, but an isopropyl substitution allowed for utilization of the H95/Y96/Q99 pocket for stronger ligand-protein interactions via van der Waal forces, resulting in a 10-fold increase in potency [70]. MRTX849 uses a tetrahydropyridopyrimidine backbone and modifications of this backbone enable hydrogen bonding with L16, H95, and G10 [69]. JDQ443 uses a pyrazole-backbone and was developed through a combination of structure-based design, optimization, and structure-activity relationship exploration. JDQ443 differentially interacts with the switch II pocket by forming hydrogen bonds with D69 and interacting with the hydrophobic pocket formed by V103, I100, M72, and Q99 [71]. MRTX849 (adagrasib) and AMG510 (sotorasib) have since shown clinical benefit in patients with KRAS^G12C-mutated non-small cell lung cancer (NSCLC) with objective response rates of 42.9% and 28.1% and median duration of responses of 8.5 and 8.6 months for adagrasib and sotorasib, respectively [72, 73]. Both inhibitors are now FDA approved for this disease.

In 2012, through NMR coupled fragment-based screening, another pocket near the switch I and switch II regions was discovered [74, 75]. BI-2852 was developed to bind within this pocket, leading to steric hindrance of the guanine-nucleotide exchange factor, SOS1 [76]. This method differs from the Shokat lab’s strategy, as this was not dependent on disulphide tethering of the G12C. This discovery led to the identification of another site near the switch II region, which, in turn, aided in the development of other inhibitors (BI-2483, BI-0474, BI-2865, BI-1823911) with in vivo activity against KRAS^G12C and pan-KRAS [77–79]. Of note, BI-2865 is a pan-KRAS inhibitor that binds with a greater affinity to the inactive conformation of KRAS regardless of the KRAS allele, but only leads to inhibition of downstream signaling in mutated KRAS [77]. Additionally, BI-18239111 is a KRAS^G12C-specific inhibitor that favors binding to the GDP-bound state and is currently in early clinical trials with the SOS1 inhibitor, BI-1701963, which shifts KRAS to the GDP-bound state (NCT04973163).

Although these therapeutic developments marked milestone in targeting KRAS, G12 is infrequently mutated to cysteine in PDAC (~1% of KRAS-mutated PDAC). However, a Phase I/II trial evaluating the efficacy of sotorasib as monotherapy in 38 patients with KRAS^G12C-mutated PDAC showed tumor shrinkage in 30/38 (79%) patients, though only 8 of those patients (21% of total participants) had a confirmed partial response [80]. The median progression-free survival (mPFS) in this study was 4 months and the median overall survival (mOS) was 6.9 months. Additionally, glecirasib (JAB-21822) was presented at ASCO GI and showed a confirmed ORR of 41.9% and mPFS of 7.0 months without reaching mOS in 31 PDAC patients [81]. Adverse effects were tolerable with both drugs, with the most frequent grade 3 treatment-related adverse effect being diarrhea with sotorasib (16%) and anemia (6%) with glecirasib. Thus, despite the low prevalence of KRAS^G12C in PDAC, G12C inhibitors can provide benefit to a subset of patients with PDAC.

From the structure of MRTX849, MRTX1133 was developed as a non-covalent inhibitor of KRAS^G12D. This was accomplished by synthesizing compounds that could interact with the negatively charged aspartate residue at position 12 [82]. Biochemical evaluation of MRTX1133 showed a 700-fold increase in selective binding of KRAS^G12D over KRAS^WT and demonstrated the ability of MRTX1133 to bind to the GTP-bound form of KRASG12D, thereby impairing downstream binding of effector proteins [83]. Testing MRTX1133 in immunocompetent mouse models of PDAC revealed that KRAS^G12D inhibition impacted cell-intrinsic and cell-extrinsic features of the tumor, mimicking genetic ablation of KRAS. Tumors cells exhibited decreased tumor cell proliferation and increased tumor cell death, while the TME showed increased myCAFs, endothelial cells, and T-cell infiltration [84, 85]. Given the efficacy in preclinical models, a phase I/II study is currently underway evaluating MRTX1133 in KRAS^G12D-mutated advanced solid tumors (NCT05737706). Given that KRAS^G12D is the most prevalent KRAS mutation in pancreatic adenocarcinoma, allele-specific targeting of KRAS^G12D has the potential to benefit many patients.

In addition to inhibitors targeting the allosteric pocket near the switch II region, tri-complex KRAS inhibitors are mechanistically distinct molecules with pre-clinical activity [86–88]. These inhibitors aim to target the active form of KRAS by forming a trimeric complex consisting of the active KRAS protein, the inhibitory molecule, and the abundant chaperone protein cyclophilin A (CYPA). Mechanistically, the inhibitor serves as a “molecular glue” by forming a binary complex with CYPA, which subsequently engages with the active form of KRAS to form the trimeric complex. Recent implementation of this strategy has resulted in the development of two molecules (RMC-6291 and RMC-4998) capable of covalently inhibiting KRAS^G12C to impair effector binding and induce tumor regression in cell-line and patient-derived xenograft (PDX) models bearing KRAS^G12C-mutated NSCLC or colorectal cancer [89]. RMC-6291 is currently in clinical trials for KRAS^G12C-mutated solid tumors (NCT05462717). Given that this strategy targets the active form of KRAS, the approach is being adapted to create allele-agnostic inhibitors. The KRAS^Multi (ON) inhibitor, RMC-6239, was shown to inhibit growth of cell-line and patient derived tumors bearing G12D, G12R, or G12V mutations. Moreover, recent clinical testing revealed activity in two patients, one with NSCLC and one with PDAC [88]. Another KRAS^Multi (ON) inhibitor, RMC-7977 was shown to impact a wide array of murine and human PDAC samples, harboring various KRAS alleles [86, 90]. Importantly, although RMC-7977 can inhibit KRAS^WT and thus impact normal tissue, it does not induce measurable cell death in normal tissues. Broadening direct inhibition of KRAS through allele-agnostic strategies will enable targeting of more tumors. Once allele-agnostic inhibitors reach further development, the adverse effects, if any, of targeting KRAS^WT in normal tissues will be of notable clinical interest.

Indirect KRAS inhibition in PDAC

Despite the limited clinical benefit seen with indirect inhibitors, circuitously targeting KRAS may still represent a viable strategy to augment treatment of KRAS-mutated tumors, particularly through combination therapy. The Src homology region 2 domain-containing phosphatase-2 (SHP2) contains two SH2 domains that recognize the phosphorylated tyrosine at position 32 within the switch I region of KRAS and can activate SOS1 to regulate KRAS-mediated signaling [91]. Phosphorylation of Y32 on KRAS impairs GTP cycling and RAS-effector binding [92]. SHP2 is important in KRAS-mediated tumorigenesis as genetic deletion of SHP2 in a KRAS mutated mouse model of PDAC decreases carcinogenesis [93]. Moreover, combination of SHP2 inhibition with concurrent KRAS inhibition demonstrates improved efficacy in preclinical models and showed TME remodeling [94]. RMC-4550 and INO155 are SHP2 inhibitors that have shown preclinical activity and are currently in early clinical trials (NCT04916236 and NCT03114319, respectively) [95, 96]. Moreover, combined inhibition of SHP2 and KRAS has been shown to overcome resistance from MET-amplification in the setting of KRAS^G12C inhibition [97].

In addition to SHP2, SOS1 has also garnered interest as a target for KRAS-addicted tumors. SOS1 activates KRAS by promoting nucleotide exchange and is also the target of negative feedback regulation by ERK-1/2-mediated phosphorylation [35]. Not surprisingly, SOS1 is required for tumor growth of implanted human pancreatic cell lines [98]. As such, the SOS1 inhibitor, BI-3406, was discovered through a high-throughput screen and was shown to work synergistically with MEK inhibitors to constrain KRAS-driven tumors, presumably by decreasing feedback reactivation in the setting of MEK inhibition [99]. Based on the structure of BI-3406, MRTX0902 was derived by screening molecules with a similar backbone to BI-3406, followed by subsequent modifications [100]. Moreover, MRTX0902 has shown the ability to limit tumor growth of a KRAS^G12C-mutated PDAC cell line (MIA PaCa-2) and appears to have synergistic activity with adagrasib. MRTX0209 is being studied in early clinical trials as monotherapy for tumors with activated MAPK pathways or in combination with adagrasib for KRAS^G12C-mutated tumors (NCT05578092). These SOS1 and SHP2 inhibitor studies suggest that utilizing a second inhibitor within the context of KRAS inhibition can help to potentiate therapeutic benefit and is reminiscent of combined BRAF/MEK inhibition used in melanoma [101].

Additionally, mitogenic signals drive cell cycle progression. However, enhanced oncogenic signaling can trigger cellular senescence through p16^INK4A or p15^INK4B encoded by CDKN2A or CDKN2B, respectively. Oncogene-induced senescence serves as a barrier to tumorigenesis [102]. However, mutation of CDKN2A impairs cell cycle arrest through accumulation of CDK4/6. The frequent co-mutation of KRAS and CDKN2A in PDAC suggest a strategy for combination of MAPK and CDK4/6 inhibition [6]. As such, preclinical models of MEK inhibition and CDK4/6 inhibition show synergistic effects, resulting in induced tumor cell senescence. These senescent cells secrete an array of cytokines to remodel the vascular microenvironment, leading to immune cell infiltration and drug influx into the tumor [103]. In organoid models, dual inhibition of ERK and CDK4/6 resulted in synergistic effects on apoptosis [104]. In a small case series of PDAC patients, 6 patients with metastatic PDAC were treated with combined MEK inhibition (Trametinib) and CDK4/6 inhibition (Palbociclib) [105]. Notably, this combination showed prolonged activity in two patients, who demonstrated clinical benefit for 9 and 17.5 months. Extrapolation of these data suggest that dual CDK4/6 and KRAS inhibition can lead to a more pronounced effect on tumor cells warrant further clinical investigation.

Immune-based strategies to target KRAS in PDAC

Lastly, immune-based therapies are being harnessed to target KRAS in PDAC. Mutant-KRAS-derived peptides are neoepitopes that can be presented on the surface of cells through allele-specific MHC molecules and recognized as tumor antigens by T cells via T-cell receptors (TCRs) [106]. Additionally, dendritic cells (DCs) infiltrate tumors and phagocytose tumor debris, allowing for processing of tumor antigens. DCs then deliver and cross-present tumor antigens to T lymphocytes in lymph nodes and other lymphoid organs to stimulate anti-tumor immunity based on the cytokine-environment [107]. These features initiate anti-tumor immunity but are eventually impaired by an immunosuppressive TME.

Multiple strategies are being employed to improve anti-tumor immunity against the KRAS antigen. Several vaccine-based therapies have shown some benefit in PDAC. A Phase I/II study in 2001, showed that intradermal injection of synthetic KRAS peptides could elicit immune responses in humans and is associated with prolonged survival in a small group of patients [108]. More recently, peptide-based immunization against mutant-KRAS have entered early phase trials and have shown KRAS-specific T-cells post vaccination in 87% of patients (NCT05726864) [109]. Other vaccination strategies, such as using listeria toxin to stimulate T-cell immunity, coupled with expression of KRAS^G12D, diminishes the progression of early PanINs to PDAC in murine models [110]. This strategy is currently being coupled with immunotherapy in the peri-operative setting of PDAC and colorectal cancer (NCT04117087) [111]. mRNA-based and mature dendritic cell vaccines against mutant-KRAS are in clinical trials (NCT03948763 and NCT03592888). The process of DC vaccination requires collection of mononuclear blood cells and ex-vivo maturation of DCs followed by exposure to mutant-KRAS peptides to produce KRAS-presenting DCs that are subsequently infused into the patient. In addition to using vaccines to stimulate the immune system, gene-therapy techniques to create tumor-targeting T-cells are being developed. A case report demonstrated that adoptive transfer of engineered autologous HLC-C*08:02-restricted TCR T-cells targeting KRAS^G12D was able to produce tumor regression in a patient with refractory metastatic pancreatic cancer [112]. As expected, there are early trials that are evaluating autologous TCR-engineered T cells for treating specific KRAS mutations in PDAC (NCT03745326, NCT03190941 and NCT06043713).

Future Challenges and Directions

Although KRAS inhibition shows clinical activity in patients, the absolute effect is modest. In CodeBreaK-200, sotorasib was compared to docetaxel in the second-line setting for NSCLC and demonstrated no OS benefit and a modest PFS benefit of 1.1 months [73]. Clinical benefit is curtailed by the development of secondary resistance, as with many targeted-therapy strategies for clinical oncology. Extrapolating from the clinical experience of KRAS^G12C inhibition in NSCLC and CRC, nearly half of acquired resistance occurs through a second-site mutation or amplification of KRAS, alternative activation of the MAPK or PI3K pathway, or through oncogenic gene-fusions [113, 114]. As such, preclinical studies are evaluating combined therapies that synergize with KRAS inhibition to overcome resistance mechanisms. These observations help drive the rationale for combination strategies in active clinical trials (Table 1). In addition to acquired resistance, 5 out of 38 patients with PDAC progressed at the time of first assessment in the Phase I/II study of sotorasib. Moreover, in Codebreak-200, nearly 15% of patients progressed on sotorasib after two months of therapy. These observations suggest that a small percentage of patients may have intrinsic resistance to KRAS inhibition or develop acquired resistance rapidly [73, 80]. Exploring these underlying mechanisms will help to elucidate better ways of improving therapy.

Table 1.

Active KRAS-directed therapy trials in Pancreas Cancer as of January 2024

Therapy	Target	Molecular details	Clinical Trial Number	Setting	Phase	Status
BI-1823911	KRASG12C	Covalently to C12 residue, locks into GDP-bound form IC50: ~10 nM (by viability)	NCT04973163	Metastatic; monotherapy or with BI 1701963 (SOS1 inhibitor)	I	Active, not recruiting
BPI-421286	KRASG12C	Covalently to C12 residue IC50: ~2.9 nM (by pERK)	NCT05315180	Metastatic; monotherapy	I	Recruiting
HBI-2438	KRASG12C	Undisclosed	NCT05485974	Metastatic; monotherapy	I	Recruiting
JAB-21822	KRASG12C	Covalently to C12 residue IC50: ~200 nM (by growth inhibition)	NCT05288205	Metastatic; given with JAB-3312 (SHP2 inhibitor)	I/II	Recruiting
LY3537982	KRASG12C	Covalent inhibitor IC50: ~3.35 nM (by viability in NSCLC)	NCT04956640	Metastatic; monotherapy and multiple combination therapy	I	Recruiting
MRTX849	KRASG12C	Covalently to G12C, affinity for GDP-bound form IC50: ~4.8 nM (by pERK)	NCT05634525	Metastatic; monotherapy	Ib	Recruiting
MRTX849	KRASG12C		NCT03785249	Metastatic; given with either pembrolizumab, Cetuximab, or afatinib	Ib	Recruiting
RMC-6291	KRASG12C	Inhibitory tri-complex with KRASG12C(ON) and CYPA IC50: ~0.11 nM	NCT05462717	Metastatic; monotherapy	I	Recruiting
RMC-6291	KRASG12C		NCT06128551	Metastatic; given with RMC-6236 (RAS-MULTI(ON) inhibitor)	I	Recruiting
INCB161734	KRASG12D	Binds both GDP/GTP forms of KRASG12D IC50: ~7 nM (by pERK)	NCT06179160	Metastatic; monotherapy or with cetuximab or retifanlimab	I	Recruiting
MRTX1133	KRASG12D	Non-covalent to switch II binding pocket IC50: ~5 nM	NCT05737706	Metastatic; monotherapy	I/II	Recruiting
RMC-9805	KRASG12D	Covalent Inhibitory tri-complex with KRASG12D(ON) and CYPA IC50: ~5nM (by viability)	NCT06040541	Metastatic; monotherapy	I/IB	Recruiting
RMC-6236	RAS-MULTI(ON)	Inhibitory tri-complex with RAS(ON) and CYPA IC50: ~1–27 nM (by viability)	NCT05379985	Metastatic	I	Recruiting
YL-17231	Pan-KRAS	Switch II pocket binding IC50: ~4–10 nM	NCT06078800	Metastatic; monotherapy	I	Recruiting
BMS-986466	SHP2	Allosteric inhibitor IC50: N.R	NCT06024174	Metastatic; given with adagrasib +/− cetuximab	I/II	Recruiting
HBI-2376	SHP2	Allosteric inhibitor IC50: N.R.	NCT05163028	Metastatic; monotherapy	I	Recruiting
JAB-3312	SHP2	Allosteric inhibitor IC50: ~3.62 nM (by pERK in NSCLC)	NCT04121286	Metastatic; monotherapy	I	Recruiting
RMC-4630	SHP2	Allosteric inhibitor IC50: ~32 nM (by proliferation in NSCLC)	NCT04916236	Metastatic; given with LY3214996 (ERK Inhibitor)	I	Recruiting
KO-2806	Farnesyl transferase	Inhibitor IC50: ~1 – 10 nM (by viability in HNSCC)	NCT06026410	Metastatic; monotherapy	I	Recruiting
Vemurafenib	BRAF	BRAFV600E inhibitor IC50: ~31 nM	NCT05068752	Metastatic; given with Sorafenib	II	Recruiting
Avutometinib	RAF/MEK	Allosteric inhibitor that forms inactive complex with RAF + MEK IC50: ~160 nM (for MEK)	NCT05669482	Metastatic; given with defactinib (FAK inhibitor) and gemcitabine/nab-paclitaxel	I/II	Recruiting
Binimetinib	MEK	Reversible inhibitor IC50: ~12nM	NCT04132505	Metastatic; given with hydroxychloroquine	I	Recruiting
Binimetinib	MEK		NCT05554367	Metastatic; given with Palbociclib	II	Recruiting
Cobimetinib	MEK	Reversible inhibitor IC50: ~4.2nM	NCT04214418	Metastatic; given with Hydroxychloroquine and atezolizumab	I/II	Active, not recruiting
BMF-219	Menin	Covalent inhibitor to disrupt Menin/Myc IC50: N.R.	NCT05631574	Metastatic; monotherapy	I	Recruiting
KRAS peptide Vaccine	Mutant KRAS	Mixture of 6 synthetic long Peptide	NCT04117087	Adjuvant; given with ipilimumab and nivolumab	I	Recruiting
KRAS peptide Vaccine (TG01)	Mutant KRAS	Mixture of 7 synthetic peptide 17 a.a. long peptides	NCT06015724	Metastatic; given with daratumumab and nivolumab	II	Recruiting
Protein Vaccine (KISIMA)	KRASG12D; KRASG12V	Recombinant fusion protein consisting of a cell-penetrating peptide, an antigen domain, and a TLR agonist domain	NCT05846516	Metastatic; given with ezabenlimab (PD-1 inhibitor)	I	Recruiting
KRAS Vaccine (ELI-002 2P)	KRASG12D KRASG12R	Lymph node targeting mKRAS-peptide and CpG DNA	NCT04853017	Adjuvant, biomarker positive, CT scan negative	I	Active, not recruiting
KRAS Vaccine (ELI-002 7P)	KRASG12D, KRASG12R, KRASG12V, KRASG12A, KRASG12C, KRASG12S, KRASG13D	Lymph node targeting mKRAS-peptide and CpG DNA	NCT05726864	Adjuvant, biomarker positive, CT scan negative	I	Recruiting
Dendritic Cell Vaccine	KRASG12D, KRASG12V, KRASG12R, KRASG12C	Autologous, peptide-pulsed, dendritic cell infusion	NCT03592888	Adjuvant; HLA-A02, HLA-A03, HLA-A11, HLA-B07 and HLA-C08.	I	Active, not recruiting
TCR	KRASG12V	Autologous, KRAS-specific-TCR transduced, T-cell infusion	NCT04146298	Metastatic; HLA-A*11:01	I/II	Recruiting
TCR (FH-A11KRASG12V-TCR)	KRASG12V	Autologous, KRAS-specific-TCR transduced, T-cell infusion	NCT06043713	Metastatic; HLA-A*11:01	I	Recruiting
TCR	KRASG12V, KRASG12D	Autologous, KRAS-specific-TCR transduced, T-cell infusion	NCT05438667	Metastatic; HLA -A*11:01	I	Recruiting
Murine TCR	KRASG12V	Murine TCR transduced, PBL infusion	NCT03190941	Metastatic; HLA-A*11:01	I/II	Recruiting
Murine TCR	KRASG12D	Murine TCR transduced, PBL infusion	NCT03745326	Metastatic; HLA-A*11:01	I/II	Recruiting

TCR – T-cell Receptor. HLA – human leukocyte antigens. IC50 – half-maximal inhibitory concentration. NR – not reported.

In addition to overcoming resistance, better pre-clinical modeling of PDAC can help define important features that predict clinical response. One limitation to pre-clinical mouse models is their isogenic nature compared to the heterogeneity of tumors in PDAC patients. Patient-derived organoids (PDOs) maintain intratumoral heterogeneity similar to their original tumors [115]. A prior study using a wide library of PDOs exemplifies the ability to discover gene-expression signatures to predict response to chemotherapy [116]. Although traditional organoid culturing conditions do not sustain stroma, recent advances to organoid technology has enabled culturing PDOs on a microfluidic chip that can also support co-culture with stromal components, allowing for further dissection of TME and tumorigenesis and therapeutic sensitivity [117]. Thus, utilization of PDOs will be a critical tool for discovering biomarkers of therapeutic response and can serve as a platform to explore new treatment vulnerabilities in the context on KRAS inhibition.

Preclinical studies demonstrate that KRAS helps to drive an immunosuppressive environment and that inhibition of KRAS results in infiltration of T-cells within tumors. Depletion of CD4+ and CD8+ T-cells in the setting of KRAS inhibition results in more rapid relapses and greater tumor growth while the addition of anti-CTLA-4 and anti-PD-1 prolongs survival in mouse in models of PDAC [84, 85]. These two observations highlight the importance of T-cells to potentiate the effect of KRAS inhibition in PDAC and raises the question of whether patients will benefit from combined immune-checkpoint blockade and KRAS inhibition.

Furthermore, the intimate relationship between extrinsic KRAS signaling and macrophage infiltration and function may be leveraged. In the pre-clinical setting, macrophage-directed therapies have been shown to sensitize PDAC tumors to chemotherapy, immunotherapy, and radiation [118, 119]. Despite this, early clinical trials with macrophage-directed therapy (anti-CSF1R antibody) showed limited clinical activity [120]. Whether synergistic effects KRAS inhibition and macrophage-directed therapy exist remains to be explored.

As we look towards these future challenges, it is important to remember that the developmental of these inhibitors would not have been possible without understanding the biochemical structure of KRAS. As such, the structure of many direct inhibitors complexed with a variety of mutant-KRAS is publicly available in the Protein Data Bank (Table 2) and may serve as a resource for the optimizing future inhibitors.

Table 2.

Publicly available co-structures with KRAS and an inhibitory molecule

PDB ID	KRAS mutant	Nucleotide	Inhibitory molecule
8AZV	KRASWT	GDP	BI-2865
8TBF	KRASWT	GNP	RMC-7977 + CypA
5F2E	KRASG12C	GDP	ARS-583
5V9U	KRASG12C	GDP	ARS-1620
6B0V	KRASG12C	GDP	ARS-107
6B0Y	KRASG12C	GDP	ARS-917
6OIM	KRASG12C	GDP	AMG 510
6USX	KRASG12C	GDP	MRTX849
5YXZ	KRASG12C	GDP	JBI484
5YY1	KRASG12C	GDP	JBI739
7R0M	KRASG12C	GDP	JDQ443
8AZX	KRASG12C	GDP	BI-2865
8G9P	KRASG12C	GNP	RMC-4998 + CypA
8TBK	KRASG12C	GNP	RMC-7977 + CypA
8JGD	KRASG12C	GDP	YK-8S
8DNI	KRASG12C	GDP	Araxes WO2020/028706A1 compound I-1
7ACA	KRASG12D	GPPCP	BI-5747
7T47	KRASG12D	GPPCP	MRTX1133
8TBL	KRASG12D	GNP	RMC-7977 + CypA
8TBJ	KRASG12R	GNP	RMC-7977 + CypA
8TBN	KRASG12S	GNP	RMC-7977 + CypA
8AZZ	KRASG12V	GDP	BI-2865
8TBM	KRASG12V	GNP	RMC-7977 + CypA
8B00	KRASG13D	GDP	BI-2865
8ONV	KRASG13D	GDP	BI-2493

CypA – Cyclophilin A. GDP – guanosine diphosphate. GNP – phosphoaminophosphonic acid-guanylate ester. GPPCP – Guanosine-5′-[(β,γ)-methyleno]triphosphate.

Conclusion

Pancreatic adenocarcinoma continues to be a fatal disease and its rising incidence is a concerning public health issue. However, decades of research have enabled greater understanding of the molecular underpinnings of this disease and have revolutionized our understanding of the role of mutant-KRAS in tumorigenesis and in the tumor microenvironment. This foundational knowledge is being transformed into therapeutic advancements that have the capacity to change clinical outcomes for patients with PDAC. Although the prognosis of PDAC remains dismal, KRAS-directed therapies represent an approach that may alter the treatment landscape of this disease.