Harnessing the power of cancer-associated fibroblasts to revolutionize pancreatic cancer treatment

FAP+ CAFs

FAP+ CAFs are predominantly localized in poorly differentiated tumor regions and have been implicated in tumor progression and immune suppression¹³. These FAP+ CAFs contribute to an immunosuppressive TME that is characterized by restricted infiltration of CD8+ T cells¹³. Depletion of FAP+ CAFs has been shown to remodel the ECM, which in turn enhances T-cell infiltration and alleviates T-cell suppression²⁰. Depletion of FAP+ CAFs has been shown to suppress tumor growth in KPC mice (an important model for PC) in an immune cell-dependent manner. Furthermore, combining depletion of FAP+ CAF with programmed death ligand 1 (PD-L1) or CTLA-4 blockade resulted in synergistic tumor growth inhibition. FAP-IL2v, an antibody against FAP and IL-2 variant, significantly improved the anti-tumor ability of PD-L1 checkpoint inhibition. The triple combination of FAP-IL2v, an anti-PD-L1 antibody, and an excitatory CD40 antibody resulted in excellent anti-tumor activity²¹. Based on preclinical research, FAP-specific chimeric antigen receptor T (CAR-T) cells, which were designed to deplete FAP+ CAFs, effectively inhibited the growth of multiple types of subcutaneously transplanted tumors by increasing endogenous CD8+ T-cell antitumor responses and inhibiting the recruitment of myeloid-derived suppressor cells (MDSCs) without significant off-tumor toxicity²². Prior administration of FAP-targeted CAR-T cells significantly potentiated the anti-PC efficacy of CLDN18.2-targeted CAR-T-cells²² and dual-targeted FAP/CLDN 18.2 CAR-T cells therapy alleviated T-cell exhaustion in a TGF-β-dependent manner²³. meso^FAP CAR-TEAM cells were generated with anti-mesothelin chimeric antigen receptor and a secreted T-cell conjugated molecule that target FAP+ CAFs and recruit T-cells via CD3. meso^FAP CAR-TEAM cells demonstrated superior efficacy in eradicating PCs and CAFs²⁴ in a PC mouse model compared to T cells targeting either antigen alone. FAP-based DNA vaccines also eliminate CAFs. The SynCon FAP DNA vaccine, a novel type of tumor antigen-specific vaccine, not only stimulates anti-tumor immunity but also shows synergistic anti-tumor efficacy with other DNA vaccines (Figure 1A)²⁵.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Therapeutic strategies for targeting CAFs in tumors. This schematic illustrates therapeutic strategies for targeting CAFs in tumors, encompassing the following three main modalities: direct depletion of CAFs; alteration of the CAF phenotype; and targeting of CAF-mediated matrix remodelling. (A) Depletion of specific CAF subpopulations. Current CAF-directed therapies can be classified into three principal categories (vaccines, cell therapy, and specific antibodies). The most advanced vaccine platforms are currently DNA-based, such as the SynCon FAP DNA vaccine, which is designed to elicit immune responses against FAP. CAR-T therapy is the predominant cell-based strategy. Primary targets for CAF-directed CAR-T therapies include mesothelin and FAP. Representative investigational agents include FAP-targeted CAR-T cells, mesothelin-targeted CAR-T cells, and meso^FAP CAR-TEAM cells. Additional strategies targeting CAF subsets include nintedanib, FAP-IL2v, anti-mesothelin antibodies, and LM10. Depletion of specific CAF subpopulations, such as αSMA⁺/LRRC15⁺ CAFs, is also under investigation. (B) Alteration of the CAF activation status. This strategy seeks to deactivate or reprogram pro-tumor CAFs rather than eliminate pro-tumor CAFs. Pharmacologic agents, such as all-trans retinoic acid, vitamin D, and the analogues, can revert activated CAFs to a quiescent phenotype. Potential agents designed to convert pro-tumor CAFs into anti-tumor CAFs include the synthetic retinoid, AM80, the heat shock protein inhibitor, XL888, and N-cadherin analogues. (C) Targeting CAF-derived ECM. This approach, which aims to remodel the ECM to improve drug delivery and alleviate immunosuppression, consists of five main strategies (inhibition of MMPs, suppression of HA synthesis, inhibition of FAK, disruption of intercellular junctions, and inhibition of collagen synthesis). Representative agents targeting MMPs include anti-MMP-7 antibodies, MC-T-DOX, and MRPL-PFD. Inhibitors of HA synthesis include PEGPH20, losartan, and minnelide. FAK inhibitors, such as VS-4718 and defactinib, represent another therapeutic category. Therapeutics directed at intercellular junctions primarily involve cadherin inhibition, as illustrated by SD133. Halofuginone is a representative inhibitor of collagen synthesis. CAFs, cancer-associated fibroblasts; FAP, fibroblast activation protein; CAR-T, chimeric antigen receptor T-cell; ATRA, all-trans retinoic acid; MMPs, matrix metalloproteinases; FAK, focal adhesion kinase; TDO2, tryptophan 2,3-dioxygenase; αSMA, alpha-smooth muscle actin; LRRC15, leucine-rich repeat containing 15; ECM, extracellular matrix; HA, hyaluronic acid. Figures were created with BioRender software (@biorender.com).

Other options

SMA+ CAFs are abundant in resected samples from patients with non-recurrent PC²⁶. However, another study revealed that decreased αSMA+ CAFs correlated with poorly differentiated tumors and poor prognosis^13,27. Although αSMA+ CAF depletion did not improve the therapeutic efficacy of gemcitabine, αSMA+ CAF depletion increased the effectiveness of anti-CTLA-4 immunotherapy and prolonged animal survival²⁷. A higher proportion of LRRC15+ myCAFs in the TME is associated with poor prognosis¹⁶. Because LRRC15+ CAFs impede CD8+ T cell effector potential, depletion of LRRC15+ CAFs from the PC stroma increases responsiveness to anti-PD-L1 treatment⁸. A subset of myCAFs that exclusively express tryptophan 2,3-dioxygenase (TDO2) promotes the conversion of CD4+ T cells into Tregs and compromises the activity of CD8+ T cells. The TDO2 inhibitor, LM10, can restore the T-cell anti-tumor response and inhibit the aggressiveness of the cancer⁷. Depletion of PDGFRβ+ CAFs with nintedanib (a multikinase inhibitor) significantly decreased the IL-6 concentration and enhanced the tumor killing efficacy of NK cells. Therefore, a therapeutic strategy combining MSLN-targeted CAR-NK cells and nintedanib may efficaciously ameliorate the clinical outcomes of patients with stroma-rich cancers by modulating the TME²⁸. Targeting mesothelin-CAFs also shows promise in the treatment of PC. Anti-mesothelin antibodies effectively reduce the Treg/CD8⁺ T-cell proportion and inhibit the mesothelial cell-to-apCAF transition, thereby improving immunotherapeutic outcomes⁹. Proton therapy can significantly increase mesothelin expression in tumors. The combination of proton therapy with mesothelin-targeting CAR-T-cell therapy in a flank PC model resulted in superior antitumor growth effects and longer survival than monotherapy. Furthermore, combination treatment in bilateral tumor models can also induce abscopal effects by increasing serum IFN-γ levels and promoting CAR-T cell proliferation (Figure 1A)²⁹.

All-trans retinoic acid (ATRA)

Activated CAFs form a physical barrier to immune cell accessibility, whereas ATRA allows stromal reprogramming through reprogramming of CAFs into quiescent fibroblasts³¹. ATRA has been shown to reduce cell proliferation and migration via the Wnt-β-catenin signaling pathway³². ATRA as a standalone treatment does not affect the progression of PC. However, when combined with gemcitabine and other chemotherapy drugs, ATRA significantly inhibits tumor growth. A phase Ib trial successfully established the safety and tolerability of ATRA in combination with gemcitabine and nab-paclitaxel in patients with PC. This treatment strategy showed promise in increasing the concentration of intratumoral chemotherapy agents while reducing adverse drug reactions³¹. Notably, ATRA treatment has also been demonstrated to increase T-cell infiltration, resulting in prolonged survival in KPC mice.

Vitamin D and its derivatives

Studies have shown that vitamin D (VD) receptor (VDR) expression is very high in CAFs. VDR is the principal regulator of CAFs. VDR ligands can enhance the chemotherapeutic response by inducing stromal remodeling. Multiple studies have shown that VD and VD derivatives promote the conversion of activated CAFs to a quiescent state. However, Gorchs³³ reported that VD has the following dual effects: VD decreases the release of pro-tumorigenic substances, such as prostaglandin E2, IL-6, and periostin, in iCAFs; and VD reduces T-cell mediated tumor immune surveillance. A clinical study involving 68 patients revealed that higher plasma VD levels were associated with longer progression-free survival (PFS) and VD supplementation may be beneficial for the prognosis of PC in patients through suppression of CAFs³⁴. The VD analogue, seocalcitol (EB1089), alone did not show objective anti-tumor activity in advanced PC in a phase II trial³⁵. A phase II pilot trial combining the VD analogue, paricalcitol, with cisplatin, gemcitabine, albumin bound paclitaxel, and nivolumab showed promising results for patients with metastatic PC³⁶.

Other options

Another study involving 71 human PCs revealed that infiltration of Meflin+ CAFs was positively correlated with favorable patient outcomes and Meflin ablation resulted in significant tumor progression with poorly differentiated histology, whereas Meflin overexpression suppressed tumor growth in mouse models. Thus, Meflin+ CAFs are regarded as anti-tumor CAFs¹⁹. The synthetic non-natural retinoic acid, AM80, can convert Meflin− CAFs to Meflin+ CAFs. Whether AM80 enhances the efficacy of cancer drugs against advanced PC is undergoing evaluation in an open-label phase I/II clinical trial³⁷. Zhang³⁰ reported that the mechanical binding of N-cadherin and HAVDI (an N-cadherin ligand) leads to transformation of activated CAFs to a quiescent state, providing a foundation for the development of novel therapeutic approaches. Treatment of primary patient CAFs with the Hsp90 inhibitor, XL888, effectively inhibits the iCAF phenotype by inhibiting JAK/STAT activity. Combined therapy with XL888 and anti-programmed death receptor 1 (PD-1) significantly increased CD8+ T-cell infiltration and prolonged the survival time of C57BL/6 mice bearing syngeneic subcutaneous or orthotopic tumors (Figure 1B)³⁸.

Type I collagen

αSMA+ CAFs are major producers of type I collagen (Col1) in the PC matrix³⁹. The Col1 synthesis inhibitor, halofuginone, significantly affects the TME in PC. Specifically, halofuginone leads to enhanced infiltration of the immune system into areas characterized by low levels of HA. Consequently, this infiltration results in a greater abundance and wider dispersion of both classically activated inflammatory macrophages and cytotoxic T cells⁴⁰. Deletion of Col1 in myCAFs was shown to accelerate PC progression and decrease survival in a dual-recombinase genetic mouse model of spontaneous PC by promoting MDSC recruitment and impeding the infiltration of CD8+ T cells, which are attenuated by the combined inhibition of CXCR2 and CCR2³⁹.

HA

HA produced by CAFs is a major source of intratumor stromal pressure. HA synthase 1 and some collagens are expressed at higher levels in iCAFs, indicating that HA synthase 1 and some collagens have specific roles in synthesis of the ECM¹². Most of the available clinical studies involving PC have focused on the effects of PEGPH20 in combination with cytotoxic drugs. A phase II study (HALO 202) reported that the combination of PEGPH20 plus gemcitabine and nab-paclitaxel significantly improved PFS in patients with PC and high HA expression⁴¹. However, a further phase III trial (HALO 109-301) revealed that this regimen did not improve the primary endpoint, overall survival (OS)⁴². A preclinical study indicated that minnelide, a water-soluble prodrug of triptolide (an active compound from a Chinese herb) not only reduced HA and collagen in mouse models but also enhanced vascular function and drug delivery within the tumor. The synergistic effect between minnelide and conventional chemotherapy not only significantly reduced the dosage of these toxic drugs but also effectively enhanced the therapeutic outcome against cancer and the stromal components⁴³. In addition, several ongoing clinical studies are further investigating the potential of minnelide. The angiotensin inhibitor, losartan, has been shown to reduce the production of interstitial HA. Losartan, in combination with FOLFIRINOX and chemoradiation, led to an improved OS for patients with PC by decreasing the number of Tregs and increasing the number of CD8+ T cells⁴⁴.

Focal adhesion kinase (FAK)

FAK is associated with ECM stiffness in PC. FAK activity in CAFs is an independent predictor of poor prognosis and tumor-infiltrating cytotoxic T cells. The FAK inhibitor, VS-4718, diminishes ECM remodeling, increases responsiveness to chemotherapy, and enhances sensitivity to immunotherapy. Notably, combining the FAK inhibitor, VS-4718, with checkpoint immunotherapy and radiotherapy or chemotherapy was reported to result in tumor eradication⁴⁵. A phase 1b/2 study confirmed the safety and efficacy of an FAK inhibitor (defactinib) and the RAF/MEK clamp inhibitor (avutometinib) in combination with gemcitabine and nab-paclitaxel as first-line treatments for metastatic PC⁴⁶.

Matrix metalloproteinases (MMPs)

Activated MMPs contribute to ECM degradation, thus overcoming the physical limitations of cell movement, which is involved in tumor invasion. MMPs may be preferable antitumor targets. Upregulation of membrane type 1-matrix metalloproteinase (MT1-MMP) contributes to increased resistance to gemcitabine in PC and MC-T-DOX is a synthetic liposome designed for tumor-targeted drug delivery. MC-T-DOX enhances intratumoral vascular density upon activation by MT1-MMP, thereby improving drug penetration and accumulation within the tumor⁴⁷. Anti-MMP-7 also increases the sensitivity to chemotherapy, which enhances apoptosis of cancer cells⁴⁸. However, the results of clinical studies on MMP targeting have not shown benefits for PC^49,50. A pirfenidone-loaded, MMP2-responsive peptide-hybrid liposome (MRPL-PFD) has been exploited for drug delivery purposes and tumors treated with therapeutic MRPL-PFD were shown to have better drug penetration and lower Col1 and fibronectin levels. Importantly, the tumor volume was greatly reduced⁵¹.

One study revealed that loss or inhibition of cadherin 11 resulted in a significant decrease in the expression of ECM components. SD133, a small molecule inhibitor of cadherin 11, effectively attenuated tumor growth and prolonged survival in KPC mice during the treatment period (Figure 1C)⁵².

IL-1/NF-κB signaling

IL-1 expression, which is associated with poor survival in PC patients, critically activates IRAK4 in CAFs, leading to fibrosis, metastasis, chemoresistance, and immunosuppression⁵³. A study (NCT02021422) is being launched involving anakinra, a human IL-1 receptor antagonist, in combination with standard chemotherapy for treating PC. Conditional deletion or pharmacologic inhibition of IRAK4 reduces the levels of immunosuppressive cytokine expression (IL-6, IL-8, CXCL2, and CXCL5)⁵⁴, reduces NF-κB activity, decreases tumor desmoplasia, increases the activity of infiltrating CD4+ and CD8+ T cells; an antitumor phenotype of CAFs was observed⁵⁴. NF-κB signaling modulates IL-1-induced IL-6 secretion⁵⁵ and is a key regulator of the acquisition and maintenance of the tumor-promoting functions of iCAFs. The gene expression profile of iCAFs was attenuated after treatment with an NF-κB small-molecule inhibitor^15,56. Garg⁵⁷ demonstrated that co-injection of orthotopically implanted KPC tumors with fibroblasts with a deletion of the p50 subunit of NF-κB reduced tumor volume and prolonged animal survival by increasing cytotoxic T-cell tumor infiltration.

LIF signaling

LIF overexpression drives iCAF phenotype via activation of the JAK-STAT signaling cascade. A clinic study revealed that the LIF protein is associated with poorly differentiated tumors. Pharmacologic blockade or genetic deletion of LIF inhibited tumor progression and enhanced the efficacy of chemotherapy in PC mouse models⁵⁸. A phase I dose escalation trial demonstrated that the LIF monoclonal antibody, MSC-1, has favorable safety and efficacy profiles in patients with advanced solid tumors. Furthermore, the unique attributes of MSC-1 in modulating the TME suggest promising opportunities for synergistic combinations with additional therapeutic agents⁵⁹. JAK inhibition resulted in a reduction in tumor volume and an increase in the number of myCAFs in KPC mice. Ruxolitinib, a JAK1 and JAK2 inhibitor, has also undergone clinical evaluation for efficacy in treating PC. Although ruxolitinib improved OS in the randomized phase II RECAP study, ruxolitinib failed to prolong OS in two randomized phase III studies⁶⁰.

IL-6/STAT3 signaling

Elevated serum IL-6 levels are a predictor of poor survival in patients with PC. IL-6 is a key regulator of STAT3 activation in cancer cells and is a primary driver of tumor cell survival and resistance to treatment⁶¹. IL-6 interacts with membrane-bound IL-6R and activates JAK/STAT3 via gp130. IL-6 deletion in CAFs was reported to modulate the TME in genetically engineered mouse models of PC, improve chemotherapy efficacy, and synergize with checkpoint blockade therapy²⁸. A trial using an anti-IL-6R monoclonal antibody (RoActemra) demonstrated significant suppression of PC progression in vivo⁶². Enhanced tumor regression and increased OS were detected in mice treated with anti-IL6Ralpha antibody and gemcitabine⁶¹. Raloxifene not only directly binds to gp130 and forms a complex with the IL-6 receptor to suppress the JAK/STAT3 pathway but also inhibits IL-6 synthesis, thereby inhibiting PC progression in vitro and in orthotopic PC xenografts⁶³. Because IL-6 targets the tumor-immune interface of PC, combined IL-6 and PD-L1 blockade elicits significant antitumor activity and increases OS in orthotopic mouse xenograft models of PC⁶⁴. The addition of MEK and STAT3 inhibitors to PD-1 blockade not only improves OS in PKT mice (a genetically-engineered mouse model of PC in which cells expressing αSMA are ablated owing to the induction of thymidine kinase by ganciclovir administration) but enhances the cytotoxic effect of T cells in the TME by attenuating myCAFs. Importantly, the combination of an MEK inhibitor (trametinib), a STAT3 inhibitor (ruxolitinib), and a PD-1 inhibitor (nivolumab) provide clinical benefits for patients with chemotherapy-resistant PC⁶⁵.

TGF-β signaling

Blockade of TGF-β signaling demonstrated the therapeutic efficacy against CAFs. The combination of the TGF-β receptor I kinase inhibitor, galunisertib, with durvalumab demonstrated good tolerability in patients with PC treated with ≤2 systemic regimens in a single-arm, multinational, phase Ib study. However, the clinical activity warrants further study⁶⁶.

CXCL12/CXCR4 signaling

FAP+ CAFs are the principal source of CXCL12 in the TME of patients with PC. Activation of the CXCL12/CXCR4 signaling pathway not only promotes the progression and angiogenesis of PC but also mediates immune escape. Preclinical studies revealed that CXCL12/CXCR4 signaling pathway blockade converts the TME from “cold” to “hot” by enhancing T-cell infiltration⁶⁰. Administration of the CXCR4 inhibitor, AMD3100, acts synergistically with immunotherapy antibodies to enhance the antitumor effects. Among 29 patients with PC who received the CXCR4 antagonist, BL-8040, and pembrolizumab as second-line therapy, the disease control rate and median OS were 34.5% and 7.5 months, respectively⁵. The combination of BL-8040, pembrolizumab, and chemotherapy was shown to be safe and well-tolerated and achieved therapeutic benefit in patients with de novo metastatic PC and disease progression on front-line gemcitabine-based therapy in the COMBAT/KEYNOTE-202 trial⁶⁷. A phase I/II clinical trial demonstrated that the CXCL12 inhibitor, NOX-A12, combined with pembrolizumab is safe and increases immune cell infiltration in the PC TME⁶⁸. These results indicated that blocking the expression of CXCL12/CXCR4 and PD-1 may increase the anti-tumor effect of chemotherapy.

HH signaling

HH signaling is aberrantly expressed in CAFs⁶⁹. HH inhibition modulates the TME by affecting ECM-related gene expression, downregulating cytokine secretion by CAFs, and altering CAF subtypes. HH inhibition using genetic or pharmacologic approaches in KPC mice, such as IPI-926 and patched 1-interacting peptide, leads to fibroblast depletion, suppresses fibrosis, reduces survival time, decreases tumor differentiation and increased vascularity⁷⁰. Conversely, activation of HH using agonists increases CAF proliferation, promotes stromal hyperplasia, and attenuates tumor growth⁷¹. Another study demonstrated that the HH inhibitor, sonidegib (LDE225), enhances the sensitivity of PC tumors to chemotherapy in mouse models⁷². HH depletion does not enhance the efficacy of gemcitabine in the treatment of PC but effectively improves tumor sensitivity to VEGFR inhibitors or anti-CTLA4 immunotherapy²⁷. Catenacci reported that vismodegib neither enhanced drug delivery nor treatment efficacy in a randomized phase Ib/II study involving patients with metastatic PC gemcitabine plus placebo or the HH inhibitor, vismodegibin and addition of vismodegib to gemcitabine did not improve the objective response rate (ORR), progression-free survival (PFS), or OS⁶⁹.

Insulin-like growth factor 1 (IGF1) signaling

HH upregulates the secretion of IGF1 and GAS6 in myCAFs, which activate the respective receptors, the IGF1 receptor (IGF-1R), and AXL, thereby activating AKT signaling. Pharmacologic inhibition of IGF-1R and AXL in vitro reverses the pro-tumor phenotypes of myCAFs but inhibition of IGF-1R or AXL only has a limited effect on decreasing tumor burden⁷³. However, the selective AXL kinase inhibitor, BGB324, enhances the efficacy of gemcitabine by modulating the immunologic landscape⁶. A recent study revealed that an AXL inhibitor (TP-0903) exhibits antitumor properties and enhances the effectiveness of other therapies in preclinical models of PC⁷⁴. TP-0903 is currently undergoing clinical trials for solid tumors (NCT02729298). A randomized, phase I/II study that focused on the safety, tolerability, and outcomes of the IGF-1R antagonist, MK-0646, in combination with gemcitabine for advanced PC demonstrated that the combination of MK-0646 plus gemcitabine was tolerable and improved OS but not PFS⁷⁵. Ganitumab, a monoclonal antibody targeting IGF-1R, showed acceptable toxicity and demonstrated potential clinical efficacy in a randomized phase II study. However, the phase III randomized, double-blind, placebo-controlled trial assessing ganitumab in combination with gemcitabine was terminated because the primary analysis indicated a negative outcome⁷⁶. In an international, randomized, double-blind, placebo-controlled phase II study involving untreated metastatic PC, adding istiratumab, an IGF1R and ErbB3 bispecific antibody, to standard chemotherapy failed to improve the ORR and OS⁷⁷. A phase Ib/II study revealed that combining the IGF-1R inhibitor, cixutumumab, with erlotinib and gemcitabine in treatment-naïve patients with metastatic PC failed to improve patient survival⁷⁸.

HGF/c-MET signaling

Overexpression of HGF/c-MET in CAFs stimulates PC cell growth and is associated with a poor prognosis⁷⁹. Treatment with the HGF neutralizing antibody, rilotumumab (AMG102), reduced the volume of the PC in mouse models and demonstrated efficacy comparable to conventional chemotherapy in impeding tumor growth⁸⁰. Crizotinib, an MET inhibitor, has been shown to prevent peritoneal metastasis of PC⁸¹. The c-MET antibody, emibetuzumab, in combination with erlotinib proved to be a safe treatment for PC in a phase I trial⁸². A combined approach using HGF/c-MET inhibitors and chemotherapy effectively decreases tumor burden and eliminates metastasis⁸³. A phase I clinical trial investigated the effect of the c-Met inhibitor, cabozantinib, on the efficacy of gemcitabine for treating PC but the sample size was small and continuous adverse reactions occurred⁸⁴. These findings provided a strong platform for the assessment of this triple therapy approach in the clinical setting⁸⁵. Another study has indicated that blocking HGF, c-MET, and urokinase-type plasminogen activator (uPA) can diminish the angiogenic properties of endothelial cells, underscoring the influence of CAFs and the HGF/c-MET pathway on neoangiogenesis⁸⁶. Further studies on the role of c-Met inhibition in PC may be needed.

Other pathways

CAFs can accelerate the invasion and migration of PC through the chemokine-receptor axis⁸⁷. Blocking the CXCL/CXCR2 axis results in alterations in the TME, which is characterized by reduced infiltration of neutrophils, MDSCs, and arginase-1+ macrophages⁸⁷. CXCL3/CXCR2 signaling can promote the transformation of CAFs to myCAFs⁸⁸. A study demonstrated that the CXCR2/1 small-molecule antagonist, SCH-479833, has both anti-tumor and anti-metastatic effects in mouse models⁸⁹. Another study revealed that inhibition of CXCR2 reversed tumor progression promoted by Col1 deletion in a PC mouse model³⁹. CCL26, which can be induced by nab-paclitaxel in iCAFs, can enhance the invasiveness of PC through the PI3K/AKT/mTOR pathway and blocking the PI3K/AKT pathway can reverse CCL26-induced invasion and migration in PC⁹⁰. CCL2 secreted by CAFs creates an immune-suppressive TME. A phase Ib study revealed that PF-04136309, an oral small-molecule CCR2 (CCL2 receptor) inhibitor, increased the sensitivity of PC to FOLFIRINOX chemotherapy⁹¹ but failed to increase the efficacy of nab-paclitaxel/gemcitabine in another phase Ib study (Figure 2)⁹².

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Targeting key signaling pathways in CAFs. Targeting key signaling pathways that mediate the bidirectional crosstalk between CAFs and tumor cells represents a promising therapeutic strategy. These functional interactions are facilitated through multiple molecular axes, including the IGF-1, HH, IL-1/NF-κB, TGF-β, CCL2/CCR2, LIF, HGF/c-MET, IL-6/STAT3, CXCL12/CXCR4, and CXCL/CXCR2 signaling pathways, which promote tumor cell survival, proliferation, migration, and CAF activation. Corresponding inhibitors against these pathways have now been developed and are under investigation in clinical or preclinical trials. These inhibitors include istiratumab, cixutumumab, ganitumab, MK-0646, BGB324, and TP-0903 against IGF-1 signaling; IPI-926, sonidegib (LDE225), vismodegib, and a patched 1-interacting peptide targeting HH signaling; anakinra, which blocks IL-1/NF-κB signaling; galunisertib, which modulates TGF-β signaling; PF-04136309, which targets the CCL2/CCR2 axis; MSC-1, an inhibitor of LIF signaling; rilotumumab (AMG102), cabozantinib, crizotinib, and emibetuzumab, which target the HGF/c-MET axis; RoActemra and raloxifene, which inhibit IL-6/STAT3 signaling; NOX-A12, AMD3100, and BL-8040, which target CXCL12/CXCR4 signaling; and SCH-479833, which inhibits the CXCL/CXCR2 axis. CAFs, cancer-associated fibroblasts; IGF1, insulin-like growth factor 1; HH, Hedgehog; LIF, leukaemia inhibitory factor; TGF-β, transforming growth factor β; HGF, hepatocyte growth factor. Figures were created with BioRender software (@biorender.com).

Metabolic reprogramming

Metabolic reprogramming is one feature of the acquisition of tumor-promoting function by CAFs. Recent experiments have shown that CAFs affect the metabolism of cancer. Cancer cells undergo a process known as the “Warburg effect” in the TME, where cancer cells convert pyruvate into lactate when there is sufficient oxygen. The excess lactic acid produced by cancer cells is transported to the TME via the lactic acid transporter monocarboxylic acid transporter (MCT) 4, which is utilized by CAFs via MCT1, thereby eliminating lactic acid from the TME^93,94. CAFs can directly feed cancer cells by secreting lactate and pyruvate in a manner that relies on the “reverse Warburg effect.” Cancer cells transform CAFs into factories that produce energy-rich metabolites for tumor progression in this scenario. CAFs undergo metabolic reprogramming and engage in glycolysis under hypoxic conditions, in which CAFs export lactate to tumor cells for oxidative phosphorylation⁹³. LDH is highly expressed in CAFs⁹⁵. The expression of LDH, which regulates lactate generation, has been identified as a poor prognostic factor in PC patients. Niftimox, an inhibitor of LDH, augments the responsiveness of PC to chemotherapy and immunotherapy via inhibition of the JAK1/STAT1 pathway and suppression of the recruitment and function of CXCR2+ neutrophils⁹⁶. Depletion of LDH A inhibits tumor growth. Lactate-stimulated CAFs upregulate IL-6 expression and cooperate with lactate to suppress cytotoxic immune cell activity. The LDH A inhibitor, FX11, reduces the tumor growth rate and increases the abundance of CD8+ T and NK cells expressing granzyme B and IFN-γ in a CAF-rich murine PC model⁹⁷. The TGF-β-SMAD5 axis promotes protein turnover flux in iCAFs and proteins secreted by CAFs support biological processes leading to cancer progression¹⁸. Studies have also demonstrated that TGF-β1 accumulation accelerates aerobic glycolysis and promotes the transport of lactic acid out of CAFs via MCT4⁹⁸. The metabolic crosstalk between CAFs and tumor cells is interrupted when MCT1 or MCT4 is blocked and pharmacologic targeting of MCT1 or MCT4 is being pursued as an anticancer therapy. AZD3965, a small-molecule inhibitor of MCT1, resulted in lactate accumulation and significant tumor growth inhibition in a Raji Burkitt’s lymphoma model⁹⁹. In a multicenter, phase I dose-escalation and dose-expansion trial, AZD3965 also showed safety and tolerability in patients with advanced solid tumors¹⁰⁰. Shikonin enhances PC sensitivity of PC to gemcitabine via suppressing reverse Warburg effect in CAFs¹⁰¹. Syrosingopine, a dual MCT1 and MCT4 inhibitor, results in glycolytic blockade, intracellular lactic acid accumulation, and synthetic lethality in tumor cells¹⁰².

Caveolin-1 (CAV1) serves as a biomarker for the “reverse Warburg effect.” CAV1 deficiency in CAFs has been shown to increase the levels of glycolytic enzymes. A reduction in CAV1 in fibroblasts is associated with activation of TGF-β signaling, which leads to the intracellular accumulation of alpha-ketoglutaric acid through inhibition of isocitrate dehydrogenase 1 (IDH1)¹⁰³. IDH1 promotes PC resistance to chemotherapy by promoting mitochondrial function and the production of alpha-ketoglutaric acid and NADPH to neutralize reactive oxygen species. Anti-IDH1 therapy combined with conventional therapy has demonstrated efficacy in treating PC in vitro and in vivo¹⁰⁴. A trial focusing on the safety and efficacy of the IDH1 inhibitor, ivosidenib, in combination with modified FOLFIRINOX is ongoing in PC (NCT05209074).

Metabolic stress caused by hypoxia, nutrient deficiency, and a stiffened stroma promotes autophagy, which promotes tumor growth in a variety of ways, such as promoting the secretion of nucleosides by CAFs that facilitate glucose metabolism and growth by cancer cells¹⁰⁵. Netrin G1 (NetG1)-expressing CAFs secrete nutrients, such as glutamate and glutamine, which can be utilized by PC cells under regulation by NGL-1¹⁰⁶. Bai and colleagues¹⁰⁷ reported that autophagy activates CAFs by promoting proline biosynthesis and collagen production. Inhibiting mitophagy by targeting PRKN, a crucial enzyme that regulates mitochondrial autophagy, reduces the tumor burden¹⁰⁷. The autophagy inhibitor, chloroquine, can effectively enhance the therapeutic effect of gemcitabine on PC³. Cancer-stimulated CAFs secrete alanine through the transporter, SLC1A4, in an autophagy-dependent manner, whereas cancer cells upregulate the transporter, SLC38A2, to transfer alanine into the cells as an alternative carbon source. Targeting SLC38A2 leads to an intratumor redox crisis and inhibits tumor growth in PC mouse models¹⁰⁸.

CAFs also promote tumor progression by providing lipids. Lysophosphatidylcholines (LPCs) secreted by CAFs are crucial components of cancer cell membranes. In addition, cancer cells secrete autotaxin, a lysophospholipase enzyme, which hydrolyses LPC into lysophosphatidic acid, promoting PC proliferation and migration through the AKT signaling pathway. These lipid-rich CAFs are then able to supply lipids to cancer cells for mitochondrial oxidative phosphorylation through the ABCA8a transporter¹⁰⁹. CAFs can facilitate communication between cells or between cells and the ECM through the production of S-type lectins. Galectin-1, a specific lectin, has a role in inducing apoptosis in T cells by binding to CD7 and CD45 present on the T-cell surface¹¹⁰. Furthermore, galectin-1 is highly expressed in regulatory T cells and is involved in the immunosuppressive function¹¹⁰. The Na+/H+ exchanger, NHE1, as the major acid extruder in CAFs, has a crucial role in the maintenance of myCAFs phenotype in a harsh acidic TME⁹⁴. Inhibition of NHE1 on the membrane of CAFs with cariporide (a selective inhibitor of the NHE1 protein) decreases the proportion of myCAFs in PCs. This effect leads to a reduced desmoplastic reaction. Adjuvant PC therapy with an NHE1 inhibitor reduces desmoplasia, shifting the immune cell infiltration from a largely innate immune cell-rich state to a more lymphocytic infiltration state (Figure 3)⁹⁴.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Targeting hypoxia and metabolic reprogramming. CAFs support metabolic reprogramming in tumor cells under hypoxic and nutrient-deprived conditions. CAFs can provide nutrients to cancer cells through various mechanisms, thereby promoting their survival. Key mechanisms include hypoxia and aberrant CAV1 expression, which promote ROS production and subsequently induce HIF activation. Accumulated HIF facilitates tumor growth via VEGF production and supplies cancer cells with abundant metabolic substrates through metabolic reprogramming, such as the reverse Warburg effect and autophagy, thereby nourishing the tumor. Moreover, acidosis within CAFs can regulate cancer cell metabolism. Under hypoxic conditions, cancer cells undergo metabolic reprogramming and produce metabolites, such as lactate, which in turn influence CAF metabolism and modulate the CAF phenotype. Strategies targeting hypoxia and metabolic reprogramming have emerged as promising therapeutic strategies. This approach can be categorized into the following three principal classes: agents targeting hypoxia, including the ROS synthesis inhibitor, resveratrol, the hypoxic microenvironment-targeting prodrug, evofosfamide, and the angiogenesis modulator, elpamotide; compounds targeting metabolic reprogramming, such as the reverse Warburg effect inhibitor, shikonin, the autophagy inhibitor, chloroquine, the lactate dehydrogenase inhibitors, niftimox and FX11, the Ref-1 inhibitor, APX2009, and the isocitrate dehydrogenase 1 inhibitor, ivosidenib; and inhibitors targeting metabolite transport, such as the MCT inhibitors, syrosingopine and AZD3965, and the NHE1 inhibitor, cariporide. CAFs, cancer-associated fibroblasts; CAV1, caveolin-1; HIF, hypoxia-inducible factor; NHE1, Na⁺/H⁺ exchanger 1; MCT, monocarboxylic acid transporter; ROS, reactive oxygen species. Figures were created with BioRender software (@biorender.com).

Hypoxic TME

Intratumoral hypoxia is a typical hallmark of PC. Recent discoveries demonstrated that hypoxia promotes tumor aggressivity and therapeutic resistance. Hypoxia signaling is stabilized by hypoxia-inducible factors (HIFs). Increased ROS production by hypoxic cancer cells induces oxidative stress in CAFs and promotes autophagy and HIF-1α stabilization. High levels of hypoxic ROS are essential for HIF-1α stabilization. HIF-1α stabilization in CAFs has a key role in promoting an inflammatory phenotype. Resveratrol significantly inhibits the hypoxia-stimulated production of ROS and HIF-1α in a concentration-dependent manner, thus inhibiting the progression of PC. N-acetylcysteine, a scavenger of ROS, inhibits hypoxia-driven ROS-induced cancer progression¹¹¹. CAF-secreted metabolites fuel the biosynthetic pathways of cancer cells¹¹². Hypoxia promotes the secretion of high levels of miR-21 extracellular vesicles by CAFs through the HIF-1α/miR-21 axis and miR-21 extracellular vesicles trigger the maintenance of PC stemness and gemcitabine resistance via the RAS/AKT/ERK pathway¹¹³. HIF2 might also be a target for myCAFs, as evidenced by a study that reported deletion of HIF2 in myCAFs dramatically decreases the intratumoral recruitment of Treg cells and immune-suppressive M2 macrophages. Treatment with the therapeutic HIF2 inhibitor, PT2399, significantly decreased in vitro macrophage chemotaxis and M2 polarization in PC animal models and enhanced the therapeutic efficacy of immunotherapy in syngeneic PC mouse models¹.

Targeting hypoxic areas with novel therapeutic approaches may offer additional anti-tumor activity and clinical benefits beyond conventional treatments. Evofosfamide is a prodrug that can be activated in hypoxic environments and induce tumor cell death. An animal study revealed that evofosfamide not only kills hypoxic PC cells but also enhances the efficacy of radiotherapy and chemotherapy¹¹⁴. The combination of evofosfamide with gemcitabine significantly improves the PFS of patients with locally advanced or metastatic PC in a randomized phase II trial compared to gemcitabine alone¹¹⁵. Systemic administration of hypoxia inducers (evofosfamide and sunitinib) to 17 patients with advanced or metastatic unresectable pancreatic neuroendocrine tumors resulted in an ORR of 17.6% and a median PFS of 10.4 months¹¹⁶.

VEGF released by hypoxic CAFs via the HIF pathway regulates pathologic angiogenesis and vascular permea-bility in patients with PC. VEGFR inhibitors have been proposed for the treatment of PC and many clinical trials have been subsequently conducted. Although many phase I trials on VEGFR have shown encouraging results¹¹⁷, the results of subsequent phase Ⅱ/Ⅲ trials have been mostly unsatisfactory. A Japanese multicenter, double-blind clinical trial used gemcitabine plus elpamotide (VEGFR 2 peptide) for PC but gemcitabine plus elpamotide failed to extend the PFS and OS¹¹⁸. Theoretically, anti-angiogenic therapy for PC can inhibit tumor proliferation and metastasis but the results are controversial (Figure 3).