The Hallmarks of Precancer

Abstract

Research on precancers, as defined as at-risk tissues and early lesions, is of high significance given the effectiveness of early intervention. We discuss the need for risk stratification to prevent overtreatment, an emphasis on the role of genetic and epigenetic aging when considering risk, and the importance of integrating macroenvironmental risk factors with molecules and cells in lesions and at-risk normal tissues for developing effective intervention and health policy strategies.

Introduction

Despite advances in screening and treatment, cancer remains a leading cause of death globally. Understanding the biology of early lesions can have major biomedical implications, exemplified by the substantial decline in cervical cancer observed after the implementation of systematic screening and preventative vaccination. However, most cancers have not experienced comparable success primarily due to a limited understanding of the precancerous state. We propose that establishing a set of generalizable principles that govern early lesions would provide insights into this crucial aspect of cancer initiation and progression. Cancers are difficult to treat due to their complex and heterogeneous nature, resulting from individualized evolutionary processes. It is widely accepted that early intervention, before the onset of complexity, is key to effectively combating cancer. Thus, identification of factors that confer risk of precancer progression and understanding their mode of action represent a critical aspect of early intervention with immediate impact. Precancerous lesions can be found in otherwise healthy individuals, as recently observed in the pancreas as well as other tissues (1), though most precancerous lesions will not progress to cancer. Consequently, employing broad, non-discriminant screening leads to cost-prohibitive over-screening and/or unnecessary treatment. Therefore, we need to explore ways to not only screen for precancers, but also perform risk stratification to identify those with the potential to advance to malignancy. Further, a deeper understanding of precancer biology will ultimately pave the way for highly effective and broadly accessible prevention and intervention strategies. The ability to identify high-risk precancers and to act early is what we envision will exert the greatest impact on reducing the societal burden of cancer.

Precancers have a broad definition, ranging from at-risk normal-appearing tissues to overt lesions from which cancers develop. While overt lesions can exhibit abnormalities in appearance, other precancerous tissues appear normal but possess molecular properties that distinguish them from both normal and cancerous tissues. Certain principles learned from cancer apply to precancers, while others do not. Consequently, many early lesions manifest with unique characteristics that are often distinct from their fully developed cancer counterparts. For instance, the metabolomes associated with cellular respiration in oral and cervical precancers are distinct from both their normal and cancerous counterparts (2). Instead, precancerous states should be viewed as discrete entities governed by their own set of rules. Aside from their broad but distinct definitions, studying precancers also presents scientific challenges. Among the most significant challenges is differentiating between advancing and non-advancing precancers, as a low percentage of precancers progress to malignancy. Additionally, tracking the trajectory of a specific precancerous lesion, especially in humans, is notoriously difficult. The clinical standard of care after detection of precancerous lesions is their complete removal, and asynchronous lesions from the same patient are genetically independent events, posing an immense challenge to determining the progression risk of any specific lesion. Strategies to partially circumvent this challenge include leveraging spatial heterogeneity within tumors to track precancerous and malignant components, utilizing medical imaging for monitoring tumor growth along with endpoint removal and characterization, and conducting large-scale epidemiological studies for finding general principles across precancer-to-cancer stages (3,4). We outline the most prevalent and influential macroscopic and molecular hallmarks across precancers derived from these studies (Fig. 1).

Figure 1. The hallmarks of precancer.

Studies on precancers have identified a set of 6 cellular/molecular traits consistently observed across various tissue types. Starting at the top center wedge of the ring and proceeding clockwise, these are: 1) age-related genetic alterations including telomere shortening and somatic mutations; 2) epigenetic changes resulting from aberrant methylation at specific sites; 3) metabolic alterations; 4) hijacking of regenerative cell state transitions; 5) disruption of immune surveillance and “inflammaging”; and 6) remodeling of the tissue microenvironment mediated by senescent fibroblasts.

Age is more than a number

By far, the predominant factor in conferring risk for developing precancerous lesions is age. Current cancer screening guidelines are primarily guided by chronological age. However, studies in specific tissues documented a wide age spectrum for the manifestation of precancerous lesions. Why the discrepancy? In understanding susceptibility and pathobiology of precancer, it is important to differentiate between chronological age and biological age. Chronological age is determined by the calendar date of birth, while biological age is based on age-associated physiological, cellular, and molecular changes observed in an individual. Biological aging can occur at a pace either faster or slower than chronological aging, and its importance is particularly pronounced in the early-onset population, which displays characteristics typically linked to advanced biological aging. Biological aging in specific tissues introduces profound modifications to the functional properties of cells, affecting both precancerous cells and those within the microenvironment by changing their genetic and epigenetic landscapes.

The primary genetic feature linked to aging in at-risk tissues was first thought to be telomere shortening, a consequence of incomplete replication of linear DNA molecule ends during repeated cell proliferation. A popular hypothesis is that telomere shortening in neoplastic cells, which activate telomerase to maintain critically short telomere lengths, may drive further genetic instability, ultimately leading to cancer progression. However, telomere shortening has also been shown to occur in the stromal cells surrounding neoplastic lesions (5), representing one of the best-appreciated genetic alterations occurring in non-neoplastic cells. When telomeres shorten to a critical length, DNA damage response pathways, including p16/Rb and p53, are activated to trigger replicative senescence in cells with inactive telomerase (6). More recent studies have shown that in addition to telomere shortening, somatic mutational burden was found to directly correlate with aging in at-risk tissues. Age-related mechanisms contribute to somatic mutations through various pathways, such as inborn errors during repeated cycles of stem cell proliferation and heightened free radical-induced DNA damage caused by the depletion of antioxidant mechanisms over time (7). Mutations in driver genes are already detected in aged “normal” tissues, and some have been found to clonally expand, covering substantial portions of the aged epithelia (8). Fifty percent or more of mutations in cancer appear to occur prior to tumor initiation (9). Mutations and copy number changes in stromal cells have also been observed to increase as a function of age (10). These studies highlight the importance of also considering molecular genetic alterations in supporting cells located in the stroma of both at-risk tissues and lesions in modulating precancer development and progression.

Another pivotal hallmark relevant to precancer risk and initiation is “epigenetic aging,” a key molecular marker that results from epigenetic drift and from a clock-like alteration in the methylation state of certain loci in the epigenome. Various clocks constructed using methylation of specific CpGs, such as the Horvath clock, are largely accurate in predicting chronological age (11). Epigenetic aging has been shown to correlate with cancer risk and has been instrumental in forecasting onset times for cancer in patients with precancerous conditions (12). The right side and the left side of the colon exhibit race-specific differences in epigenetic aging, aligning closely with the higher incidence and earlier onset of tumor development in the right colons of African American individuals (13). Age-dependent epigenetic alterations, akin to genetic changes, regulate cell states and influence disease progression.

Chronological age is unalterable, but certain aspects of biological aging can be modified. Currently in industrialized nations, modification to biological age is unfavorable due to widespread exposure of the population to macroenvironmental risk factors associated with premature aging. These often-interrelated factors include adoption of a Western diet, meat consumption, obesity, alcohol and tobacco use, adverse alterations to the microbiome, socioeconomic marginalization, and limited healthcare access. The altered metabolic environment associated with a high-sugar diet and obesity provides a specific example of how macroenvironmental factors potentially integrate into biological age (14). Western diets lead to dysregulation of blood glucose control and hyperinsulinemia. Insulin is a potent growth factor that can act on premalignant states (15), and this is thought to be responsible for the increased incidence of obesity-associated cancers, including those with close association with mutations along the insulin-PI3K signaling axis. Consumption of high fructose, a feature of the Western diet, is one factor that contributes to an earlier onset age of colon cancer (16). How adverse macroenvironmental factors contribute to the recent rise in early onset cancers is currently an intense area of research. Altered biology to induce precancer are entwined with healthcare disparities and systemic biases. Neighborhood-wide association studies have identified variables such as transportation accessibility and food deserts as significant in altering precancer incidences (17). Chronological aging, epigenetic aging, and social determinants of health can also be compounded. For example, vitamin C deficiency is linked to food insecurity and is often observed in older people. Low vitamin C can also accelerate leukaemogenesis by reducing the function of Tet2, a tumor suppressor and regulator of DNA methylation (18). Understanding the biology of precancer is thus critical not only for shaping screening and intervention strategies but also for determining public health policy.

The holy grail of precancer research with the highest impact on cancer is to understand which, how, and when macroenvironmental factors influence cellular and molecular functions, predisposing specific cells to transformation and certain precancers to progress into malignancy. Here, we discuss significant transitions in epithelial and non-epithelial cells resulting from macroenvironmental exposure, shedding light on how at-risk tissues and lesions are selectively inclined to progress towards neoplastic states.

Regenerative mechanisms driving epithelial transitions in premalignancy

Abnormal growth, a hallmark of cancer, requires cells progressing towards malignancy to transition towards a state with dysregulated cell proliferation. Applying the concept of cancer as “wounds that will not heal”, premalignant cells hijack normal pathways of regeneration to facilitate abnormal growth. Cellular damage arises from exposure to macroenvironmental risk factors. For instance, reactive oxygen species generated by altered metabolism due to overnutrition, a high-fat diet, and obesity can result in cellular DNA damage. Changes in the microbiome attributed to lifestyle differences can also injure epithelial surfaces. Signatures of genotoxicity induced by colibactin-producing bacteria (e.g. Escherichia coli encoding the pks genomic island) have been observed in colorectal and prostate precancerous tissues (19,20). Furthermore, the most prevalent microbial drivers of tumor initiation, Human papillomavirus in cervical cancer, hepatitis B virus in liver cancer, and Helicobacter pylori in gastric cancer, are associated with cellular damage and inflammation (21). While other microbe-associated factors, such as circulating microbial metabolites and cytokines generated secondarily by immune cells, can also modulate tumorigenic processes (22-24), damage repair mechanisms occurring locally appear the most dominant in shifting cell states during the early stages of premalignancy.

In rapidly renewing tissues, regeneration is carried out by tissue-resident stem cells with high proliferative and self-renewal capacities. Dysregulation of their expansion and differentiation by sequential oncogenic alterations give rise to the development of precancerous lesions and malignant progression (25-27). While replicative stress during multiple cycles of homeostatic renewal (hence, aging) may already increase the risk of stem cell transformation, chronic regeneration may further escalate oncogenic processes by accelerating stem cell proliferation. Differentiation from these abnormal stem cells generates an aberrant hierarchy of cell lineages and states that create intratumoral heterogeneity that enables evolution from precancer to cancer. The majority of tissues do not have tissue-resident stem cells and must rely on other mechanisms for damage regeneration. Mounting evidence suggests that metaplasia is a common regeneration mechanism observed frequently in epithelial tissues. Metaplasia is defined as the replacement of one mature differentiated cell type with another differentiated cell type abnormally presented in the tissue, usually in the context of damage. Intriguingly, metaplasia can arise from differentiated cells by dedifferentiation or transdifferentiation without the involvement of tissue-resident stem cells (28). Metaplasia serves as a reversible, protective mechanism that leads to tissue regeneration in response to environmental insults. However, prolonged metaplasia due to persistent exposure is also a risk factor for dysplasia and malignant transformation (29). Many processes underlying metaplasia development parallels oncogenic pathways, as exemplified by activation of autophagy, mTORC1, and p53 in paligenosis - a stepwise mechanism of differentiated cell reprogramming to a progenitor-like state following tissue damage (30). Notably, immunological cues and communication with stromal cells induce metaplastic reprogramming during tissue remodeling in response to damage (29). For example, Laminin α5 from stromal fibroblast has been shown to induce acinar-to-ductal metaplasia in the pancreas (31).

Previous studies highlight distinct subtypes of precancers arising from different routes of tissue regeneration (25,26,32). One remarkable example is the evolution of colonic polyps, where the serrated polyp subtype originates from differentiated cells through pyloric metaplasia and the conventional adenoma subtype originates from the uncontrolled expansion of tissue resident stem cells (32). It has also been shown that mechanisms of precancer cell transitions are conserved across different tissues, as pancreatic acinar-to-ductal metaplasia and Barrett’s esophagus share common features with pyloric metaplasia (33,34). Admittedly, cell states contributing to precancer initiation and progression can extend beyond stemness and metaplasia, and future studies may reveal additional routes of cellular transitions occurring in precancerous lesions.

Microenvironmental drivers of precancer initiation and progression

A vigorous immune system's ability to identify and eradicate abnormal cells is a major component in preventing tumor development. Neoantigens, which arise during the early phases of neoplasia in genetically abnormal cells, alert and activate the adaptive immune system to engage in the elimination of aberrant cells. This process, termed immunosurveillance, operates silently in the backdrop of individuals with a robust immune system, effectively preventing genetically altered cells from becoming established. Recent precancer atlases have revealed that an initial cytotoxic immune microenvironment, marked by the presence of cytotoxic T cells and natural killer cells, is a consistent feature in almost every precancerous tumor microenvironment (TME) (35). Immunosurveillance potentially curtails many precancerous lesions from progressing, with a higher risk of progression in lesions with a more immunosuppressed TME.

A positive immune response emerges through the coordinated interplay of not just adaptive immune cells but also the orchestration of other supportive cells in the microenvironment to create a harmonious symphony. For instance, ICAM-1+/VCAM-1+/PDPN+ immunofibroblasts secrete a chemokine cocktail to establish immune cell microarchitectures known as tertiary lymphoid structures, comprising primarily B cells and T cells that are structurally intercalated by immunofibroblasts and lymphatic vessels (36). These structures are found in precancer and cancer tissues and are associated with good clinical outcomes through supporting a favorable adaptive immune response.

When the initial immunosurveillance fails to eradicate the early aberrant cells, precancerous lesions can manifest and progress. The decline in adaptive immune responses is closely linked to the molecular hallmarks of aging, and this association is a key aspect of immunosenescence, which is influenced by both organismal factors such as thymic involution, and local alterations to immune cells. Changes in cellular responses, as illustrated by inflammaging, can be ascribed to molecular factors of biological aging, leading to a gradual shift from adaptive immunity to innate immunity over time. A consequence of inflammaging is chronic, low-grade inflammation driven by innate immunity, accompanied by an age-dependent increase in circulating inflammatory molecules. This baseline inflammation is also observed in other conditions such as obesity and chronic inflammatory disorders that accelerate biological age, serving as a key predisposing factor for tumor development. Apart from excessive sugar intake, reduced oxygen levels and impaired blood vessel function during the aging process can create a metabolic environment reminiscent of the Warburg effect. Alternative usage of glycolysis intermediates leads to heightened lactate production and acidity, subsequently suppressing the production of cytotoxicity factors in CD8+ T cells and NK cells. Increased lactate levels have also been shown to favor the polarization of CD4+ T cells to regulatory T cells and macrophages to a more M2 pro-tumorigenic phenotype (37). In addition, genetic and epigenetic changes occurring with age can intrinsically alter immune and stromal cells in the microenvironment leading to senescence, cell death, and dysfunction.

A major factor in driving TME dysregulation is a disrupted interaction between fibroblasts and immune cells. Precancer atlas studies have revealed an augmentation in stromal cell content and a reduction in cytotoxic T cells as precancer progresses towards malignancy (35). Age-related genetic and epigenetic alterations directly lead to fibroblast senescence, which is associated with dysregulated cellular behavior in addition to a halt in replication. The senescence-associated secretory phenotype (SASP), characterized by production of interleukins, chemokines, other pro-inflammatory factors, growth factors, and proteases, suppresses immune responses and induces precancer cell proliferation (38). Increased deposition of collagen by fibroblasts, in particular, is associated with aging and a desmoplastic TME can limit immune cell trafficking by establishing physical immune exclusion barriers (3). Emerging research has also shown that cancer associated fibroblasts (CAFs) that express many of the same markers as senescent fibroblasts modulate the conversion of the TME from an adaptive to an innate one replete with macrophages and myeloid-derived suppressive cells (39).

Senescent cells may also act in both a juxtacrine manner, and a paracrine manner via the SASP, to “spread” senescence from one cell to another (40). While precise identification of senescent cells remains challenging due to the absence of validated markers, the establishment of gene sets such as SenMayo (41) and use of proteomic analysis to generate SASP atlases (42) are major steps forward in the study of senescence. Application of these tools to the study of precancers provides an opportunity for researchers to assess the efficacy of therapies targeting senescent cells, either directly or via modulating SASP.

Future perspective into precancer research

The prevailing argument in the field is that molecular alterations are less heterogeneous in precancers compared to their cancerous counterparts, potentially offering a window for effective intervention. A potential avenue is to employ cytotoxic immunity to control precancer initiation and progression, akin to the observed success in immunotherapy applied to cancer, which could even be more effective at the precancerous stage. However, a significant challenge remains in identifying and stratifying patients based on progression risk to avoid over-diagnosis and over-treatment. Another challenge is to understand factors that can be leveraged to favorably alter the immune system, striking a balance between inflammation and cytotoxic immunity to target tumors. Our current perspective stresses the complex connection between macroenvironmental factors and biological aging, modifying the microenvironment that acts as fertile ground for precancer development. Prevention and intervention approaches should consider the interplay between these components.

It goes without saying that adopting a healthy lifestyle can mitigate the risk of chronic diseases, including cancer. However, there is a large chasm in the understanding how macroenvironmental risk factors translate into molecular and cellular changes, an essential insight for the development of therapies effective against precancer. While biological age stands out as the most significant risk factor, many current interventions aimed at enhancing "health span" lack rigorous research due to gaps in understanding of the complex connections between the macroenvironment and the microenvironment. For instance, the complex interaction between antioxidants and the initiation of cancer has complicated prevention trials aimed at targeting endogenous antioxidant pathways. The Beta-Carotene and Retinol Efficacy Trial and the Alpha-Tocopherol, Beta-Carotene Cancer Prevention, which aimed to decrease lung cancer incidence and progression, had to be stopped early because antioxidant supplementation was found to increase lung cancer incidence in smokers and people exposed to asbestos (43,44). The identified challenges and gaps in knowledge provide avenues for future research that can inform effective strategies against precancer.

The observation of cell-state shifts of key microenvironmental components underscores the importance of high-resolution studies integrating multi-omic and spatial information to accurately depict the complex process of precancer development (Fig. 2). Next generation profiling data must incorporate patient level metadata necessary to uncover the complex interconnection between the macroenvironment, microenvironment, and disease progression. Precancer research is intimately tied to cancer disparities, as diverse populations exhibit distinct genetic and socio-economic predispositions. There is an enormous need to comprehensively understand how these factors are molecularly integrated to induce biological changes in cells in order to inform effective public health policies. Innovative experimental model systems incorporating tumor cells and microenvironmental cells are needed to mechanistically validate many of these essential associations. There is also a high translational value in investigating precancers that currently do not have effective prevention strategies, such as precursors to bladder and pancreatic cancers. We look to a bright future where these multidimensional approaches and insights are poised to significantly advance our understanding of cancer development, bridging the current gap between the present day and a future where we have stamped out these devastating diseases.

Figure 2. Integration of multiomics analyses with patient metadata to understand how macroscopic risk factors influence molecular pathways.

Advancing our knowledge of the cellular and molecular interactions that contribute to the formation and progression of precancers is essential as we build towards a cancer-free world. We propose integrative analysis of patient metadata with next-generation multiomics to enhance our understanding of how macroenvironmental exposures molecularly affect cells and biological pathways.