Adult Pancreatic Acinar Progenitor-like Populations in Regeneration and Cancer

Abstract

The bulk of the pancreas is primarily composed of long-lived acinar cells that are not considered a bona fide source for stem cells. However, certain acinar subpopulations possess a repopulating capacity during regeneration, raising the hypothesis as to the presence of regenerative progenitor-like populations in the adult pancreas. Here, we describe recent discoveries based on fate-mapping techniques that support the existence of progenitor-like acinar subpopulations, including active progenitor-like cells that maintain tissue homeostasis and facultative progenitor-like cells that drive tissue regeneration. A possible link between progenitor-like acinar cells and cancer initiators is proposed. Further analysis of these cellular components is needed, as it would help uncover possible cellular sources for regeneration and cancer, as well as potential targets for therapy.

The pancreatic acinar compartment

The pancreas is a digestive organ that contains enzyme-producing acinar cells and ductal cells that facilitate exocrine secretion, as well as hormone-secreting endocrine cells. The endocrine role of the pancreas is mediated by four major Functionally distinct islet cells: α cells (glucagon), β cells (insulin), δ cells (somatostatin) and PP or γ cells (pancreatic polypeptide) [1]. Most of the volume (over 95%) of the pancreatic epithelium is made up by acinar and ductal cells. The acini and ducts are physicially connected by centroacinar cells, which are, perhaps, more of a terminal extension of the intercalated ducts [1].

In the pancreas, where research into stem and progenitor cells (see Glossary) has recently received increased attention, bona fide stem cells have yet to be defined, although there is growing recognition of progenitor-like cell diversity. Recent fate-mapping studies have uncovered an unexpected level of heterogeneity or plasticity in pancreatic acinar cells, with distinct regenerative and tumorigenic potential encompassed by subpopulations of progenitor-like cells within the exocrine compartment [2–4]. Notably, acinar cell subpopulations, which can convert to a cancer stem-like cell state [3,5], are attractive candidates for the possible cells of origin for pancreatic ductal carcinoma (PDAC) (see Clinician’s Corner), the most prevalent form of pancreatic cancer [6–8].

Clinician’s corner:

• Most attention on pancreatic regeneration has focused on the development of therapies for diabetes. Identifying and utilizing islet stem/progenitors to produce β cells for this purpose has been challenging given the limited availability of such cellular sources. To date, acinar cells have not been shown to be very promising as a source of cellular therapies for diabetes or other diseases.

• The facultative progenitor-like population, while largely quiescent under normal conditions, has been proposed to be important cellular sources for regenerative therapy [84,85]. In theory, such a cellular population might be useful to aid in the treatment of patients suffering from pancreatitis or recovering from surgical resection or after pancreatectomy.

• Regenerating the exocrine pancreas has been considered equally important because of the recognition of certain normal cellular populations residing in the exocrine compartment that may play a protective role against tumorigenesis. Investigations characterizing and targeting such populations may assist in the prevention of pancreatic cancer.

• Recurrence of pancreatic cancer after R0 resection is as high as 90%, presumably as a result of residual cancer-initiating cells. Therefore, it will be essential to identify and track these populations and to develop therapeutic strategies to target these cells.

This review focuses on acinar cell subpopulations and discusses the possible presence of stem/progenitor-like cell populations. Further, we discuss how these recently uncovered acinar cell subpopulations may be involved in tissue regeneration and cancer initiation. Whether other pancreatic epithelial compartments (such as ducts) may also serve as a source of progenitors is not the focus of this review.

Genetic tracing addresses the adult pancreatic epithelial lineages

To address questions of stem cells and cellular lineages within the adult pancreas, the CreER/loxP-based system of genetic labeling followed by in vivo tracking of daughter cells over time has been most commonly used, and has proven to be a valuable tool [9]. The CreER/loxP system employs a variant of Cre recombinase fused to a mutated human estrogen receptor (Cre-ER), allowing for tamoxifen-dependent Cre recombination and thus the labeling of cells in a time-controlled manner [10]. So far, lineage-tracing studies of the acinar epithelium have been reported using a number of promoter-driven inducible Cre(s) (reviewed in [9,11], Table 1). Analysis of these promoter-driven reporters (e.g., Mist1, Ptf1a, Elastase, Bmi1, Dclk1, etc.) has demonstrated distinct lineage-tracing patterns, with variable proportions of the acinar population traced, ranging from <1% to >90%. There are a number of limitations to these studies, including variability in the approach used for the generation of inducible lines (e.g., transgene, BAC, or knock-in), the reporter alleles followed for tracing studies (e.g., fluorescent proteins or LacZ), the type of induction regimen (e.g., tamoxifen or doxycycline), and the faithfulness of transgene expression. Even with these caveats, quantitative assessment over time strongly suggests inherent differences in progenitor-like abilities among acinar cell subpopulations. Nevertheless, the possibility cannot be completely excluded that some of the differences may result from variability in the efficiency of labeling certain acinar cells.

Table 1.

Genetic labels used to trace pancreatic acinar epithelium in adult animals

Labeled compartment	Inducible Cre lines	Labeling description	Reference
Acinar cells specific	Mist1-CreERT2 (knock-in)	Label over 90% of the acinar cells	[57,70]
	Ptf1a-CreERTM (knock-in)	Mosaic recombination in acinar cells; ~20% (label index) after a 7-day chase	[78]
	Ptf1a-CreERTM (knock-in)	Label 60%–80% of acinar cells after induction	[43]
	Elastase-CreERT (BAC)	Label 100% of acini and leakage in ~50% of adult acinar cells	[79]
	Elastase- CreERT2 (transgene)	Label 30% of acinar cells in a patchy or lobular pattern; low leakage	[41]
	Ela-CreERT (transgene)	Label >50% acini at 2 weeks after induction; <1% leakage	[60]
	proCPA1- CreERT (knock-in)	Label a maximum of 20% of acini	[80]
	Nestin- CreERT2 (transgene)	Label ~23% of the acinar cells	[4]
Acinar and centroacinar cells	Elas-tTA/tetO-Cre (knock-in)	Label 20%–30% of acinar cells and some centroacinar cells	[69]
Acinar and islet	Bmi1-CreERT(knock-in)	Label ~2.5% of the pancreas at 5 days after induction; including acinar cells and glucagon+ cells	[2]
Acinar, duct, and islet	Dclk1-CreERT2(BAC)	Label 0.1%–0.5% pancreatic epithelium including acinar cells, centroacinar, ductal cells, and rare endocrine cells	[3]
	Human Carbonic anhydrase II-CreERT (transgene)	Label patches of acinar, ductal and islet cells	[81]
Other inducible Cre lines mentioned in this review
	Sox9-CreERT2 (BAC)	Label an average of 70% of ductal cells, including terminal duct/centroacinar cells	[82]
	Sox9-CreER (transgene)	Label 12% all Sox9+ cells and 0.02% of acinar cells	[5]
	Krt19-CreERT(knock-in)	Label 10–45% of ductal cells (P0 to adult), < 1% of acinar cells and < 1% of islet cells	[46]
	Rip-CreERTM (transgene)	Label ~30% of β-cells and 3–5% acinar cells	[83]

A progenitor-like cell pool for regeneration and cancer

Over the past decades, embryonic multipotent progenitor cells (MPC) in normal pancreas development [12–16] have been identified. The early embryonic multipotent population, derived from foregut endoderm, express the transcription factors Pdx1, Prox1, Onecut-1, Ptf1a, Foxa1/2, Tcf2, Sox9, Gata4/6, and Hes1. In the late development, bi-potent progenitors give rise to duct and endocrine lineages and progenitors located in the distal pancreatic tip produce the acinar lineages [17]. However, in the adult pancreas under conditions of both homeostasis and regeneration, the presence and location of a bi- or multipotent progenitor population has been debated [13,18]. Unlike endocrine cells, the acinar compartment of the pancreas has until recently attracted limited attention, especially as a source of stem or progenitor cells. Nevertheless, fate-mapping studies have uncovered an unexpected level of heterogeneity or plasticity in pancreatic acinar cells, with distinct regenerative and tumorigenic potential encompassed by subpopulations of progenitor-like cells within the exocrine compartment [2–4]. Notably, the ability of some acinar cells to maintain or convert into a progenitor-like cell state in vitro has been shown [19,20].

Similar advances have been made in our understanding of the cellular origins, of the cancer-initiating cells (CIC) in pancreatic ductal carcinoma (PDAC), the most prevalent form of pancreatic cancer [6–8]. Several acinar cell subpopulations, mapped by inducible, cell-specific Cre- lineage-tracing experiments (such as Dclk1-Cre^ERT), are attractive candidates for the possible cells of origin for PDAC, which can convert to a cancer stem-like cell state [3,5]. Nevertheless, at present, there is no consensus as to the CICs in PDAC. It is certainly plausible that identification of CICs may not be limited to a single marker or even a single progenitor lineage. Indeed, there may exist more than one type of acinar cell subpopulation that is able to contribute to PDAC initiation. Characterization of such CIC populations would aid in the development of targeted strategies for PDAC prevention and therapy.

The acinar cell population is heterogeneous

In spite of the fact that acinar cells are morphologically indistinguishable from each other, there has long been speculation in the field that acinar cells are a heterogeneous species. Cellular heterogeneity in some cases is defined or interpreted by discrete molecular markers [21]. Emerging techniques of single-cell analysis, which permit the study of genomics, transcriptomics, proteomics and metabolomics at the single-cell level, has facilitated the identification of unique genetic markers for distinct subpopulations as well as providing some mechanistic insights [22]. In the pancreas, recent evidence from single-cell RNA sequencing (scRNAseq) analysis has demonstrated the presence of molecularly distinct acinar subpopulations [4,23]. For example, scRNAseq analysis of the human pancreas has revealed the presence of large clusters of REG3A⁺PRSS1⁺ acinar subpopulations in proximity to the islets of Langerhans [23]. However, a few recent scRNAseq studies and meta-analyses of available scRNAseq datasets have suggested that the reported heterogeneity in the pancreas may be due to inadequate sampling of the transcriptome, or to a transitory state captured at one point in time, rather than necessarily reflecting a fundamental difference between cells within the acinar population [24,25]. Clearly, further studies are needed to better assess the heterogeneity of the acinar compartment and confirm that any molecular distinctions are meaningful at a functional level.

Moreover, differences in genetic or epigenetic information may often lead to phenotypic disparities among acinar subpopulations. Indeed, lineage-tracing studies have provided strong evidence for acinar cellular and phenotypic heterogeneity. Acinar cells labeled by the Bmi1 promoter contain a subpopulation capable of self-renewal [2]. Stmn⁺ acinar cells are transiently activated and actively cycle following injury [4]. Quiescent Dclk1⁺ acinar cells serve as facultative progenitor-like cells that are crucial for pancreatic tissue repair [3]. Proliferative heterogeneity of acinar cells has also been reported using both multicolor lineage-tracing strategy and in ex vivo organoid culture [4].

Acinar cells are distinct in their ability to produce digestive enzymes [26,27]. Immunocytochemical evidence from early pancreas studies suggested differences in intracellular enzyme content between acinar cells in close proximity to the islets compared to more distantly situated acinar cells [28]. This is analogous to the liver, where hepatocytes located in the periportal region express differential gene signatures from their peri-venous hepatocyte counterparts [29,30]. There is still controversy about whether such regional differences are due to the physiological function of those cells in response to differential nutrient supply (e.g. oxygen and nutrient content in the blood), or to the greater presence of certain progenitors in the periportal area than in the peri-venous area [31]. Moreover, epidemiological analysis of PDAC prevalence demonstrated that PDACs more commonly arise from the head of the pancreas (60–70%) compared to the body/tail (20–25%) [32,33]. This raises questions as to whether histologically “geographic” distinctions exist among the pancreatic acinar subpopulations, such that (1) acinar cell subpopulation located close to centroacinar cells are distinguishable from those further from the centroacinar cells, and/or (2) acinar cell subpopulation located near lymph nodes of the pancreas head are distinct from those at tail. In addition to possible anatomic distribution, the effect from external influencers, such as microbiome, can not be excluded. It is likely that the cells or CICs in the head of the pancreas may be subjected to more exposure to the intestinal microbiome than the body/tail area. Further single-cell analysis of different anatomical regions of the pancreas using microdissection techniques may help resolve these questions and provide useful insights into the anatomic distribution of CICs for PDAC.

Acinar epithelium harbors both active and regenerative progenitor-like cells

Self-renewal of the acini is maintained by active progenitor-like cells

Cellular proliferation within the acinar population is crucial to the maintenance and renewal of acinar compartment. Houbracken and colleagues demonstrated that in the immediate postnatal period, more than 40% of the acinar compartment consists of Ki67⁺amylase⁺ cells, indicating the presence of an abundant progenitor/progenitor-like population that is actively dividing and sustaining acinar cell expansion after birth [34]. In adult homeostasis, such a Ki67 labeled population is decreased to <2% of the acinar compartment [26,34], implying that active progenitor-like cells or acinar-progenitor cells are markedly diminished after transition to adulthood.

While the existence of stem or progenitor cells in the adult pancreas has been debated [16], it has been difficult to determine whether only a small subset of acinar cells, or whether all acinar cells, retain the ability to proliferate and contribute to regeneration. The stem/progenitor model is likely a better concept to explain recent observations and thus address these questions. Indeed, long-term lineage-tracing studies have implied differential lineage behavior by progenitor-like cells in conditions of homeostasis. Bmi1⁺ and Nestin⁺ lineages both contain actively proliferating acinar subpopulations that can renew part of the acinar compartment and sustain proliferation for over one year [2,4]. Dclk1⁺ and Stmn⁺ cells are long-lived, largely quiescent and lack proliferation under resting conditions [3,4]. Although these cell populations have never been firmly termed as true progenitors, these experiments raise the possibility of active progenitor-like characteristics that might contribute to self-renewal and proliferation of the acini in homeostasis.

Regenerative progenitor-like cells contribute to tissue regeneration after injury

Although less remarkable than the liver, the pancreatic epithelium also demonstrates significant regenerative capacity in both the endocrine and exocrine compartments [35]. In adult rats, a large portion of the pancreas mass can be restored within three weeks after 90% partial pancreatectomy [36]. A complete restoration of the acinar compartment can be achieved in mice within one or two weeks after CCK analog caerulein-induced acute pancreatitis [37,38]. In spite of this clear regenerative ability, the type and degree of injury may determine the pattern of regeneration and repair. Furthermore, a key unanswered question for decades has been the cellular origins of the newly formed pancreatic epithelium. So far, two predominant models have been proposed and debated [39,40].

One is that regeneration comes from self-replication of pre-existing terminally differentiated exocrine or endocrine cells, which implies that mature cell types undergo dedifferentiation into progenitor-like cells while retaining their lineage identities. Supporting evidence comes from genetic labeling of acinar cells in mice subjected to acute injury, which showed that the acinar compartment can undergo mass regeneration from replication of Elastase-expressing acinar cells [41]. However, it is unclear whether all the Elastase-expressing acinar cells are equally capable of replication.

The second model favors the notion that acinar expansion is derived from reserve stem or progenitor cells (e.g., facultative progenitor cells) that contribute to the replacement of tissue mass [42]. The facultative progenitor cell model has been presented as a “hybrid” between stem cells and progenitor cells [42]. A number of studies have demonstrated a variable degree of contribution from the acinar compartment to the repopulation of the injured pancreatic epithelium, and have identified specific acinar subsets as being facultative progenitor-like populations [2–4]. In the classical caerulein-induced injury model, the Bmi1⁺ lineage expands more than two-fold at five days after caerulein treatment, while the Stmn-expressing acinar cells increase from 1% to more than 30% of the acinar population at four days [2,4]. In particular, quiescent Dclk1⁺ acinar cells, labeling only 0.1–0.5% of the acinar population, can repopulate the entire newly-formed pancreatic lobes after partial pancreatectomy [3].

Using a pancreatic ductal ligation (PDL) injury model to investigate possible cellular sources of β islet cells, several studies have conducted lineage-tracing in mouse lines expressing CreERT under the control of the acinar cell promoters. Pan and colleagues demonstrated that the Ptf1a⁺ acinar cell population, which comprises ~80% of acinar cells, is capable of transdifferentiating into β cells in the PDL-injured pancreas [43]. Several other studies have also shown through growth factor stimulation or genetic manipulation (e.g., co-expression of a transcription factor cocktail) that the acinar population can occasionally transdifferentiate into β cells [44,45]. However, Desai et al. showed that the Elastase-Cre^ERT2 construct, which labels 30–40 % of acinar cells, did not demonstrate that acinar cells could convert to any other lineages after partial pancreatectomy (PPx), PDL, or caerulein-induced pancreatitis [41]. Although these somewhat contradictory results may be attributed to discrepancies in experimental conditions, a possible alternative interpretation is that, rather than the entire acinar population, only a very restricted acinar subpopulation is capable of converting into endocrine cell fate under very specific conditions. However, the exact identity of such a subpopulation remains unclear.

Regenerative progenitor-like cells and cellular plasticity

Cellular plasticity of the adult pancreas is of considerable recent interest and a topic addressed in several excellent reviews [14–16]. In a number of lineage-tracing experiments, some degree of acinar cell plasticity has been demonstrated. This has now been documented in three settings: first, through in vitro organoid culture experiments, where trans-differentiation of acinar cells into ductal cells has been frequently observed; second, in vivo during the process of acinar-to-duct metaplasia (ADM) [46–48]; and third, during the rare transdifferentiation of acinar cells to β cells, hepatocytes or adipocytes [43,49,50]. However, it seems unlikely that all acinar subpopulations possess the same degree of cellular flexibility and plasticity, since most studies have shown that not all acinar specific Cre-driven lineages can efficiently form organoids in culture or give rise to other cell types [19,54]. Perhaps, rather than all acinar cells broadly exhibiting some level of plasticity, only a subset of the progenitor-like acinar cells are able to show such plasticity.

The regenerative potential is influenced by niche and intrinsic factors

Assuming regeneration of the pancreas is mostly driven by facultative progenitor-like cells, the question of how acinar progenitor-like cells engage in tissue regeneration has not been well explored. In light of recent literature, two possibilities can be proposed. One is that regeneration is triggered by microenvironmental factors that can influence the engagement of the regenerative cell populations. In adulthood, acinar progenitor-like cells are largely quiescent under resting conditions, with proliferation inhibited by local factors that also help to define acinar identity. When these inhibitory restraints are removed as, for example when acinar cells are isolated and grown in organoid culture, transdifferentiation of acinar to ductal cells is remarkably accelerated [51]. It can be assumed that ductal cells, immune cells, endothelial cells, stellate cells, nerves, bone marrow-derived cells or even surrounding acinar cells, may influence the progenitor-like activity or determine which type of progenitor-like cells will be selected to contribute to the regenerative process [52].

Another possibility is that the diverse behavior of acinar progenitor-like populations is related to their intrinsic genetic or epigenetic status, similar to the behavior observed in the hematopoietic system [53]. This has been well documented through gain- and loss-of-function studies targeted to specific genes of the acinar or epithelial compartment. For instance, whole-body knockout of Bmi1 led to impaired acinar regeneration through increased apoptosis and bone marrow deficits [54]. Interestingly, deficiency of Bmi1 may not affect short-term normal homeostasis of the pancreas [54], suggesting that facultative progenitor-like cells active in regeneration, but probably not progenitor-like cells active in homeostasis, are dependent on Bmi1. Overexpression of the important acinar cell differentiation factor Mist1 solidified acinar cell characteristics, which results in a reduced progenitor-like potential to maintain tissue homeostasis [55]. Deletion of Mist1 supports acinar cell dedifferentiation towards a progenitor-like phenotype, which then exhibit increased proliferation and sensitivity towards cerulein-induced ADM formation [56]. Loss of c-Myc in the pancreas impaired normal pancreas homeostasis and resulted in acinar-to-adipose transdifferentiation as the mice aged [50]. Disruption in acinar cells of the Xbp1 gene, an endoplasmic reticulum (ER) stress regulator, evokes robust regenerative responses from the remaining intact acinar cells [57]. These data suggest that specific transcription factors may be critical for both active and regenerative progenitor-like cells. On the other hand, deletion of β-catenin in the Elastase-Cre^ERT- or Pdx1-Cre- labeled acinar population appears not to alter the maintenance or the sensitivity of acinar cells to caerulein-induced injury [58–60]. Altogether, these studies might imply that acinar progenitor-like subpopulations for regeneration may be sensitive to the deficiency of certain transcription factors such as Mist1, Bmi1, c-Myc or Xbp1 but less so to β-catenin, while acinar progenitor-like subpopulations for homeostasis may be less sensitive to Bmi1 or β-catenin deficiency.

It may be equally reasonable to assume that along with genetic discrepancies, the cellular behaviors of acinar cell subpopulations are programmed by differential epigenetic memory. This epigenetic status may confer to progenitor-like cells distinct gene expression profiles and/or distinct capacity to respond to signaling from mutant Kras [61–63]. Recent single-cell analysis has begun to investigate the cellular heterogeneity at the epigenetic level, such as transcription factor binding, DNA methylation, and chromatin states of single cells, which will perhaps allow us to better understand the epigenetic determinants of cell fates and lineage decisions in a spatiotemporal fashion [64].

Are all acinar cells equally potent in initiating PDAC?

Targeted activation of oncogenic mutant Kras in the murine pancreatic epithelium has served as a highly relevant model for studying human PDAC. Under this oncogenic condition, and especially when targeted to the epithelium with a pancreas-specific Cre such as Pdx1 or Ptf1a, pancreatic intraepithelial neoplasia (PanIN) lesions can be initiated postnatally. These PanIN lesions show sequential progression to PDAC, and as such faithfully recapitulate the PanIN-to-PDAC sequence in human patients [65]. In the classic mouse models, oncogenic mutant Kras is activated and expressed during embryogenesis and in all pancreatic epithelial lineages. In human pancreatic cancer, however, oncogenic mutant Kras is mostly seen as a rare, stochastic mutation that arises postnatally [66]. To study potential cells-of-origin for PDAC in appropriate mouse models, extensive preclinical studies have been conducted in mice expressing a mutant Kras in various pancreatic specific lineages in adulthood. Unexpectedly, activation of mutant Kras in ductal cells (under the control of Krt19, Sox9) or islet cells (under the control of Insulin promoters, e.g., Rip) have failed to initiate PDAC in adulthood [5,67,68]. Acinar cells, instead, are found to be much more susceptible to mutant Kras transformation than duct cells (an estimated >100-fold) [5]. This observation is relevant to human disease, as 94% of pancreatic cancer develops in the exocrine tissue of the pancreas [8].

Given that acinar cells are more susceptible to Kras-dependent transformation and tumorigenesis in mouse models, the following question must be asked: do all the acinar cells harbor the same capacity of tumorigenesis? Interestingly, lineage-specific activation of mutant Kras with certain acinar-specific inducible Cre drivers (such as Mist1, Elastase, proCPA1, etc.) demonstrated a variable capacity for PanIN formation, resulting in histological changes ranging from no lesions to abundant high-grade PanINs [5,69,70]. Why acinar cell behavior varies greatly in the context of activated mutant Kras is currently unclear, and explanations have focused on inflammation and/or microbiome as the possible key factors that drive the tumorigenesis in the setting of mutant Kras [69,71,72]. Nevertheless, the likelihood that acinar subpopulations harbor unequal potency in responding to Kras signaling and thus a variable contribution to pancreatic tumorigenesis has not been fully excluded. Similar to regeneration, the likelihood that acinar subpopulations harbor unequal potency in contributing to pancreatic tumorigenesis appears to be most consistent with evidence to date. Moreover, since several PDAC molecular subtypes have been recently described [73], the question of whether distinct progenitor-like populations can give rise to different PDAC subtypes remains unanswered and warrants future studies.

Are regenerative progenitors and cancer-initiating cells linked?

Regarding the relationship between tissue regeneration and cancer, one can hypothesize that both processes are connected to the same stem or progenitor-like cells. This is an intriguing prospect and highlights the importance of determining whether regenerative lesions or cancer can emerge from dedifferentiation of all the acinar cells or derive from a specific subpopulation of the acinar progenitor-like cells. Definitive answers to these questions remain elusive at present. The notion that regenerating progenitor-like cells are also cancer-initiating has been proposed [6,74]. Indeed, Dclk1⁺ facultative progenitor-like cells are responsible for both regeneration and tumorigenesis [3]. Importantly, Dclk1⁺ cells do not efficiently initiate PanINs unless the caerulein-induced injury occurs, suggesting that even in the context of activated mutant Kras, oncogenesis of these CICs need to be triggered by injury, perhaps indicating a requirement for activation of a regenerative or repair program. On the other hand, a recent paper showed that while Sox9⁺ hybrid liver progenitors are essential for regeneration, they cannot give rise to liver cancer in three independent cancer models [75], suggesting the absence of overlap in the cellular sources responsible for liver regeneration and cancer. Whether there exist certain progenitor-like cells that might prefer the initiation of cancer as opposed to regeneration will require additional investigation and lineage-tracing studies.

Concluding Remarks

Despite the many decades of interest in identifying adult endocrine progenitors of the pancreas, with the ultimate goal of procuring a cellular source for treating diabetes [76,77], studies of adult acinar progenitor-like cells have only recently become an area of active interest. Since adult pancreatic stem cells have yet to be identified, the focus instead has been on various adult acinar populations, which likely play essential roles in pancreatic regeneration and cancer development. However, these acinar populations may comprise one or more progenitor-like subpopulations. The question of whether all acinar cells or only a subset are self-renewable, regenerative competent or oncogenic-sensitive, requires further studies (Figure 1, also see Outstanding Questions). Emerging single-cell technologies will no doubt further advance our understanding of acinar biology and cell-fate in the pancreas. In any case, characterization of such acinar progenitor-like cells will likely provide new avenues for advances in pancreatic regenerative medicine and cancer therapy.

Figure 1. Models for pancreatic homeostasis, injury and cancer initiation.

The “2-progenitor-like cells” model is also illustrated for the homeostasis and regeneration. The putative stem cell model is not included in this illustration. PDAC: pancreatic ductal adenocarcinoma. Upper panel, green: differentiated acinar cells that are considered identical in the self-replication model. Middle panel, green: active progenitor-like cells, purple: facultative progenitor-like cells, orange: non-progenitor-like cells. Lower panel: green square: a PDAC subtype originate from active progenitor-like cells, purple hexagon: a PDAC subtype originate from facultative progenitor-like cells.

Outstanding Questions Box:

• Regarding the acinar compartment, which specific acinar subpopulations harbor cellular plasticity?

• What are the molecular mechanisms that determine cell fate?

• How can we best take advantage of the cellular plasticity for use in tissue regeneration and repair?

• Since studies on molecular classification of PDAC have identified distinct PDAC subtypes, the future question is whether distinct acinar progenitor-like cells give rise to different PDAC subtypes.

• Cancer is considered as “unresolved” tissue regeneration. While in the pancreas, facultative progenitor-like cells also contribute to cancer initiation, evidence from some other organs, supports that cellular sources for regeneration or cancer are distinct. So how to resolve such distinctions in order to better harness cellular sources for tissue repair while avoiding cancer initiation?

• Finally, how can we better translate our understanding of novel stem/progenitor populations to the development of therapeutic targets and cell-based therapies for patients with pancreatic diseases?

Highlights:

• The acinar compartment in the murine pancreas is heterogeneous and comprises a pool of progenitor-like cells.

• A “two-progenitor-like cells” hypothesis is proposed for the development and maintenance of the exocrine pancreas. Active progenitor-like cells maintain the homeostasis and self-renewal while facultative progenitor-like cells contribute to repair.

• Acinar cell compartment is more susceptible to PDAC and harbor cancer-initiating/progenitor-like cells. Acinar progenitor-like cells are distinct in their ability to initiate cancer.

• Regenerative progenitor-like cells contribute to cancer in the pancreas but not in some other organs. Dissecting the relationship of regenerative progenitor-like cells and cancer-initiating cells may help better understanding and identification of the potential targets for therapy.

Glossary

Acinar-to-ductal metaplasia (ADM)

a process that pancreatic acinar cells transdifferentiate into ductal-like cells or adapt a ductal-like cell fate

Acinar-progenitor cells

Unipotent progenitor-like cells that can divide but only give rise to terminally differentiated acinar cells

Active progenitor-like population

similar to transit-amplifying progenitors described in the, this cell population may arise from the stem cells and divide a limited number of times to differentiated cells under homeostasis

Cancer-initiating cells (CIC)

In the setting of the early non-tumorigenic or preneoplastic pancreas, CIC refers to the cells which acquire the first mutations and give rise to the incipient cancer, and these often have the potential for malignant transformation (a). In the setting of tumor recurrence, CIC refers to the resident or circulating cancer cells, which remain in the body after surgery and are responsible for tumor recurrence (b)

Cellular or cell fate

a cell from its birth to division and/or differentiation to a particular cell type. Here, we describe the phenomenon of acinar cells converting into other cell types under experimental conditions

Facultative progenitor or progenitor-like cells

“demonstrate differentiated features and remains a quiescent state under normal homeostasis but in response to injury” [86]. Here, we refer to the potential facultative progenitor-like population that is quiescent in homeostatic conditions but is activated and expanded to replace tissue loss in response to injury

Multipotent progenitor cells

have the capacity to develop into more than one specialized cell type in a specific tissue or organ

Lineage tracing

is a method originally used in developmental biology to study the origin of various tissues and is now an important tool for studying stem/progenitor cells in adult tissues. In lineage tracing, a single cell is labeled and the mark is transmitted to all the cell’s progeny, resulting in labeled clones

Plasticity

a term “generically captures changes in cellular identity or phenotype that occur outside normal development and tissue homeostasis”, and commonly describes cell type or state changes in an adult tissue [87]. In this review, we focus on acinar cell plasticity during tissue regeneration

Potency

in cell biology, it commonly describes a cell’s ability to differentiate into other cell types. Here, it describes the ability of various acinar cell subpopulations to convert to pancreatic cancer precursor lesions under mutant Kras

Progenitor cells or progenitor-like cells

Compared to stem cells, progenitor cells have limited self-renewal and lesser potential to give rise to other cell types in vivo in the absence of experimental manipulation. The concept of different subclasses of progenitor cells is also evolving. Here, we loosely refer to potential progenitor-like populations that (1) are capable of converting to a progenitor state under any experimental conditions and (2) have the potential to divide and expand during homeostasis or tissue regeneration. These cells differ from progenitor cells found during embryonic development

R0 resection

microscopically pathologically complete removal of the tumor (R0)

Stem cells

distinct from differentiated cellular populations, stem cells are defined by their ability to self-renew and differentiate into other cell types [6]