Fateful fat: Intra-pancreatic lipids cause pancreatic cancer

Abstract

In a Mendelian randomization and prospective cohort study,¹ intra-pancreatic fat increases the risk of pancreatic cancer. This provides persuasive human evidence of causal relation between lipids and cancer in the pancreas, which confirms a prediction of the PANDORA hypothesis.

In a Mendelian randomization and prospective cohort study, intra-pancreatic fat increases the risk of pancreatic cancer. This provides persuasive human evidence of causal relation between lipids and cancer in the pancreas, which confirms a prediction of the PANDORA hypothesis.

Main text

Preventing, intercepting, and treating sporadic pancreatic ductal adenocarcinoma (PDAC) is notoriously difficult, at least in part because of incomplete understanding of what causes it. Conspicuous general adiposity is a well-known risk factor for PDAC, but it is non-specific, being associated with cancer in other organs (e.g., esophagus, colon, kidney, and breast) too. Inconspicuous local fat within the pancreas—intra-pancreatic fat deposition (IPFD)—might be a specific determinant of PDAC, and this was the premise of the study by Yamazaki et al.¹ in this issue of the Cell Reports Medicine.

The link between IPFD in non-cancerous pancreas tissue and PDAC—a cancer characterized by abundant desmoplastic reaction to a relatively small volume of cancer cells—was investigated in several earlier studies. A systematic review and meta-analysis showed that more than half of patients with pancreatic cancer or pre-malignant lesions had fatty change of the pancreas (i.e., disorder characterized by excess IPFD).² Moreover, these patients had a nearly 3-fold greater probability of having fatty change of the pancreas in comparison with controls.² Some studies also found that fatty change of the pancreas was significantly associated with PDAC after taking into account body mass index,^3,4 suggesting that fatty change of the pancreas is not necessarily a function of general adiposity in patients with PDAC. However, all the primary investigations in humans to date have invariably been conventional observational studies. This is a key barrier to understanding the role of IPFD in PDAC, as the direction of causality cannot be ascertained from these studies. More specifically, it is possible that PDAC causes fatty change of the pancreas (e.g., through PDAC-associated pancreatic duct obstruction leading to increased IPFD), in which case pondering about fatty change of the pancreas would have limited clinical implications.

The use of genetic variants as natural experiments—Mendelian randomization—has emerged as an acceptable way to establish causal relation between a modifiable risk factor and a disease because genetic variants are fixed at conception. This method is much less likely to be influenced by reverse causation and confounding (both measured and unmeasured) than conventional observational studies. Yamazaki et al.¹ leveraged Mendelian randomization to investigate the link between IPFD and PDAC. The authors employed a total of eight previously identified single-nucleotide polymorphisms of IPFD (below the conventional genome-wide significance p value threshold of 5 × 10⁻⁸) and found a statistically significant relation between increased IPFD and PDAC. This consistently held true in a series of sensitivity and subgroup analyses. Notably, when the three genetic variants nominally associated with body mass index were excluded, the relation remained significant (in fact, it became even stronger). The authors also used an additional study design (a multivariable Cox regression model applied to prospective cohort with a relatively long follow-up period) that yielded a similar result to the two-sample Mendelian randomization analysis and, hence, demonstrated consistency of relation between increased IPFD and PDAC across different study designs.¹ This body of work represents an important contribution, as it provides the first causal, rather than associational, evidence of the role of IPFD in the development of PDAC in humans. Ideally, such evidence would have come from a sufficiently powered randomized controlled trial. However, as the use of this study design to investigate the effect of IPFD on PDAC is not feasible in humans, the strength of evidence obtained by the authors can only be matched, but not surpassed, in the years to come.

The core findings presented by Yamazaki et al. go some way to confirm the PANDORA (PANcreatic Diseases Originating from intRa-pancreatic fAt) hypothesis.⁵ It postulates that fatty change of the pancreas is the single most important non-genetic structural abnormality that underlies all common diseases of the pancreas (such as type 2 diabetes, diabetes of the exocrine pancreas, acute pancreatitis, chronic pancreatitis, and PDAC). In particular, the fifth prediction of the PANDORA hypothesis is specifically related to increased IPFD heightening the risk of sporadic pancreatic cancer.⁵ Fatty change of the pancreas might facilitate hypoxia-driven cell reprogramming, release of proinflammatory cytokines, increase in presence of cancer cell progenitors, and result in the development of a pancreatic microenvironment advantageous to PDAC formation.^5,6 It is fitting to mention at this point that fatty change of the pancreas also predisposes to pancreatitis, which is one of the strongest known long-term risk factors for PDAC.^7,8

While the detailed cellular interactions involving lipids and carcinogenesis in the pancreas are yet to be elucidated, PDAC likely develops through the compounding effect of multiple routes (often over decades) that reflect the multipronged nature of IPFD (Figure 1). Adipocytes between the pancreatic lobules are the most obvious suspects in pancreatic carcinogenesis as retroperitoneal fat can invaginate the pancreas.^2,6 But there are several other resident cells in the pancreas, and processes involving these cells, that also contribute to IPFD and may play a role in pancreatic carcinogenesis. Both acinar cells and endocrine cells in the human pancreas can contain lipid droplets—the dynamic organelles storing fatty acids (in the form of triglycerides and esters) and directing them to various pathways, especially under hypoxic conditions.⁹ The highly plastic acinar cells can also change their identity and, for example, be turned into adipocytes through acinar-to-adipocyte transdifferentiation.⁹ Mesenchymal stem cells (located in the adventitia of pancreatic arteries and veins) underpin fatty replacement of parenchymal cells after their death by lipid peroxidation or other pathways.¹⁰ Last, pancreatic stellate cells—typically located near pancreatic ductules—are known to contain lipid droplets (that accumulate retinyl esters and palmitic acid) in the quiescent state. The loss of lipid droplets when pancreatic stellate cells are activated to cancer-associated fibroblasts may play a role in the intercellular trafficking that sets the stage for PDAC.⁹

Figure 1.

Routes to pancreatic carcinogenesis that involve intra-pancreatic fat

Elements unrelated to IPFD are omitted. Drawings of resident cells in the pancreas are not to scale. β cell is presented as an exemplar of endocrine cell in the pancreas.

Irrespective of the exact cellular mechanisms, accumulating evidence compels consideration of lipid derangements in the pancreas as the fundamental target for prevention or interception of sporadic PDAC. The integrated PANDORA model “moves the needle” on thinking why pancreatic cancer develops. The next time somebody tells you that fate throws a lot of curves at a person with pancreatic cancer, wonder whether it may actually be fat (in their pancreas).