Evolution of the MUC1 gene in eutherian mammals as an adaptation responsible for the increasing incidence of cancer in humans

Abstract

The incidence of cancer in humans has been rising in association with extended life spans. Incidence rates of early onset cancers in humans <55 years of age have also been increasing for unclear reasons. One potential contributory factor is an antagonistic pleiotropy in which certain genes that appeared in mammals to increase fitness for reproduction contribute to cancer susceptibility later in life. A related concept is an evolutionary mismatch in which humans have adapted to certain environmental and dietary factors that change over time and thereby increase cancer incidence. The MUCIN 1 (MUC1) gene emerged in mammals and represents an example of antagonistic pleiotropy and evolutionary mismatch that is posited here as a contributing factor to the increasing incidence of cancer in humans. This Review focuses on the roles of MUC1 and the oncogenic M1C protein in reproductive fitness and barrier tissue protection that in settings of chronic inflammation promote pan-cancer progression and treatment resistance. Also highlighted are therapeutic approaches targeting MUC1 and M1C that are under clinical and pre-clinical development.

1. Introduction

Cancer is found in all mammalian species; whereas incidence and mortality rates differ widely, especially in eutherians [1–3]. As formulated in Peto’s paradox, cancer mortality risk in mammals is considered largely independent of body mass and longevity [4]. In support of Peto’s paradox, a large study of mammals found no relationship between body weight and cancer death rates [3]. Concordantly, as often cited examples, long-lived large mammals, such as elephants and whales, have a low cancer risk [5]. Contrary to Peto’s paradox, recent work has reported that cancer prevalence increases with adult mass and somatic mutation rate, but decreases with gestation time [6,7]. These concepts relate to cancer in mammals in general; however, if and how they apply to the increasing prevalence of human cancers remains unclear.

The incidence of cancer in humans has been rising in association with an extended life span [2,3]. Moreover, incidence rates of early onset cancers in humans <55 years of age have been increasing, particularly for carcinomas of barrier epithelial tissues [8–14]. Early onset of 29 different carcinomas, including those of the stomach, colon and rectum, have risen worldwide by almost 80% from 1990 to 2019 [15]. Among those, colorectal cancers were the most common worldwide in 2022 [14]. The underlying causes of these cancers remain unclear, but likely relate to the increasing prevalence of chronic inflammation arising from changes in our dietary and environmental exposures [3,16–18].

Natural selection is believed to shape cancer susceptibility, prevalence and resistance [1,19,20]. As a corollary to that premise, cancer-causing genes acting later in life are not necessarily expunged from the human genome by natural selection [21]. Alleles promoting reproductive success are advantaged by selection even if they later reduce lifespan [22]. In supporting this concept of antagonistic pleiotropy, certain genes that increase fitness during reproduction can contribute to cancer susceptibility later in life [23,24]. A lack of selective pressure after reproductive age has therefore been proposed to explain increases in cancer incidence in the elderly [25]. A related concept is an evolutionary mismatch in which cancers in humans <55 years of age, as well as those older, are caused by a contradiction between our evolutionary adaptations and the prevalent increases in chronic inflammation associated with lifestyle changes [3,16–18]. Genes that evolved in mammals and contribute to antagonistic pleiotropy and/or evolutionary mismatch can, when viewed from this perspective, provide unforeseen insights into the basis for the increasing incidence of cancers in humans.

2. Evolution of the MUC1 gene in eutherian mammals

The MUCIN 1 (MUC1) gene was identified following detection of the encoded protein in human breast cancers and human milk fat globule membranes [26–29]. MUC1 includes variable numbers of 60 bp tandem repeats [30,31]. The encoded serine, threonine and proline rich 20 aa repeats are extensively modified by O-glycosylation [30,31]. This unique physical structure formed the basis for identification of a family of genetically distinct secreted and transmembrane mucins that differ by numbers of repeats and their amino acid sequences [32]. Ancestral mucin genes, such as MUC6, MUC2, MUC5AC and MUC5A, encode secreted mucins that form a physical gel in protecting barrier epithelia [33–36]. MUC1 differs from these secreted gel-forming mucins by an evolutionary adaptation in primates and hominids encompassing the introduction of a transmembrane and cytoplasmic protein to extend protection against loss of epithelial homeostasis [35–37].

MUC1 differs from other transmembrane mucins by a distinct pattern of expression restricted to therian and eutherian mammals [35,36]. Genes that arose in mammals largely include those encoding proteins that are expressed in reproductive organs, as well as other tissues [38]. For instance, MUC1 is expressed in reproductive and barrier epithelia of eutherians [39]. Another example is the widely expressed genes encoding the apolipoprotein B mRNA-editing catalytic 3 (APOBEC3; A3A, A3B, A3C, A3D, A3F, A3G and A3H) enzymes that evolved in mammals to restrict viral replication [40,41]. Coevolution of MUC1 and the A3s apparently occurred in integrating defenses against viruses, which can disrupt reproduction and maternal-fetal symbiosis, as well as maternal survival necessary for propagation of offspring [42].

MUC1 is expressed in the oviduct [43] and in spermatocytes and spermatids of the testes [44]. MUC1 is also expressed in the fetus and uterus, as well as lactating mammary gland, indicative of emergence of this gene for enhancing fitness of mammalian reproductive functions. Additionally, MUC1 evolved in mammals for the protection of barrier tissues, such as the respiratory and gastrointestinal tracts, specialized organs including the liver and kidney, as well as skin, to ensure sufficient longevity for the delivery and nurturing of offspring (Fig. 1).

Fig. 1.

Evolution of MUC1 in Eutherian Mammals. MUC1 appeared in placental mammals to enhance fitness for reproductive functions and for protection of barrier tissues to ensure propagation of offspring.

2.1. Involvement of MUC1 in placentation

Evolution of mammalian placentation required adaptive mechanisms enabling trophoblast invasion of the uterus and avoidance of fetal rejection by the maternal immune system. Regulation of MUC1 in uterine luminal epithelial cells occurs within a window for implantation during the estrous cycle [45]. Provocatively, attachment of the embryo is regulated by MUC1 expression and localization [39,46,47]. MUC1 is also expressed by the outer layer of placental syncytiotrophoblasts that invades the uterine wall during implantation [48](Fig. 2). As a result, MUC1 is expressed at the apical surface of the maternal-fetal interface (MFI) throughout pregnancy and at increased levels with gestational age [48,49].

Fig. 2.

MUC1 expression at the maternal-fetal interface. MUC1 evolved in eutherian mammals to support placentation and fitness of the fetus by pleotropic functions at the MFI. Figure adapted from [42].

The MFI at the labyrinth in mice and placental villous in humans has remained a paradox for understanding tolerance [50,51]. Maternal acceptance of a fetus expressing paternally inherited allo-antigens has supported adaptation of the immune system in response to placentation [52]. Mouse Muc1 and human MUC1 have been linked to regulation of (i) decidual macrophages and NK cells [53] and (ii) placental interactions with the innate immune system [49](Fig. 2). What has remained unclear is whether the complex roles of Muc1/MUC1 at the MFI involving inflammatory attachment and tolerance in placentation relate to immune evasion in cancer [46,52].

MUC1 evolved in therian and eutherian mammals, the latter of which give birth to more developed offspring than therians. The therian marsupials have a placenta like eutherian mammals, yet it consists of only a few layers of cells and is short-lived [46]. Term pregnancy in the therian opossum is characterized by an inflammatory response and Muc1 expression consistent with that observed in eutherian mammals [46]. This inflammatory response in therians leads to parturition; whereas, in eutherians, an adaptation involving a shift to a non-inflammatory setting allowed for an extended period of placentation [46]. One interpretation of these findings is that eutherian Muc1/MUC1 may function in protecting trophoblast-mediated invasion and damage of the endometrium. Interestingly, the Muc1/MUC1 transmembrane and cytoplasmic domains are highly conserved in therians and eutherians [35,36]; whereas, there are significant differences in their tandem repeat sequences and numbers, indicating adaptations in the extracellular domain that may have occurred to support eutherian placentation and reproductive fitness.

2.2. Production of MUC1 by the lactating mammary gland

Another distinction of eutherian mammals is a mammary gland that produces milk-fat globule membranes (MFGMs) containing high MUC1 levels [28,54]. MUC1 is expressed at the apical membrane of lactating mammary epithelial cells and functions in physically maintaining open mammary ducts and lumens [54]. Beyond providing nutrients, breast feeding is well-recognized for neonatal protection against pathogens [55]. MFGM MUC1 protects against rotavirus and E. coli infections in the neonatal gastrointestinal tract [56,57]. MUC1 also inhibits Salmonella invasion of intestinal epithelia [58]. As another example, human milk MUC1 blocks transmission of HIV-1 from dendritic cells (DCs) to T cells [59]. Furthermore, MUC1 binds to DC intercellular adhesion molecule 3-grabbing non-integrin (DC-SIGN) in regulating interactions with pathogens [55].

In summary, the MUC1 gene appeared in eutherian mammals to support placentation and fitness of offspring during lactation. MUC1 also evolved to protect barrier epithelia from the external environment to support survival of adults for bearing and nurturing progeny. MUC1 represents a notable example of antagonistic pleiotropy as a gene that promotes reproductive fitness and contributes to cancer susceptibility later in life [23,24].

3. Expression of MUC1 by barrier epithelia

The human MFI functions as a barrier between the trophoblast-lined placental villus and maternal immune system (Fig. 2). The MFI can therefore be considered a barrier tissue for placentation. By analogy, barrier tissues interfacing with the external environment are defined as compositions of (i) polarized monolayers of glandular epithelial cells, and (ii) resident lineages of stromal, hematopoietic and neural stem cells, among others. These barrier tissues line the respiratory and gastrointestinal tracts, where they balance metabolic functions with host defense [60]. Polarized epithelia lining ducts of specialized organs, such as the liver and kidney, also represent barrier tissues that afford protection by the clearance of environmental toxins. The skin is another barrier tissue composed of stratified epithelial cells that protect against biotic and abiotic insults [61]. Importantly, barrier tissues have the remarkable capability for survival by recalling insults that more effectively inform responses to subsequent exposures [60,61]. Emerging evidence signifies that this capacity for memory plays an essential role in the major challenge of resistance that cancer cells acquire during treatment.

3.1. MUC1 is expressed at the apical borders of polarized barrier epithelia

MUC1 evolved to protect barrier epithelia against loss of homeostasis [32,62]. As one line of defense, MUC1 forms a protective physical barrier, comparable to that bestowed by the ancestral secreted gel-forming mucins [33–36]. MUC1 evolved from the MUC5AC secreted mucin [35,36], which interestingly is also widely expressed in barrier epithelia and their corresponding carcinomas [63]. The region of MUC1 containing the glycosylated tandem repeats is the mucin component of the molecule. The 60 bp tandem repeats, encoding PAHGVTSAPDTRPAPGSTAP, can vary in number from 20 to 120 depending on the allele at exon 2 (Fig. 3).

Fig. 3.

Structure of the MUC1 heterodimeric protein. The MUC1 gene localized at 1q22 includes 7 exons. Exon 2, which emerged from ancestral secreted mucins, encodes the VNTR of 20 aa repeats that are modified by O-linked glycans. Exons 4–7 encode the transmembrane MUC1-C/M1C region that was incorporated as an adaptation in mammals. MUC1 is translated as a single polypeptide that undergoes auto-cleavage at the SEA domain encoded by exon 3. The MUC1-N-terminal and M1C subunits in turn form a non-covalent complex that is expressed at the apical borders of polarized barrier epithelial cells. MUC1-N is the mucin component of the complex. M1C is a non-mucin transmembrane protein.

During evolution and as an adaptation from the secreted gel-forming mucins, MUC1 acquired sequences encoding a transmembrane protein [32,35,36]. MUC1 also evolved with sequences encoding a sea urchin sperm protein, enterokinase and agrin (SEA) domain between the mucin domain and the non-mucin transmembrane protein (Fig. 3) [32,35,36]. Incorporation of SEA domain sequences was critical for MUC1 function in protecting barrier tissues [62]. In this way, MUC1 is translated as a single polypeptide that in the endoplasmic reticulum (ER) undergoes auto-proteolysis at the SEA domain [62] (Fig. 3). In turn, the resulting MUC1 N-terminal (MUC1-N) and C-terminal (MUC1-C/M1C) subunits form a non-covalent leucine-mediated complex that is transported to the epithelial apical cell membrane [62]. There and like the ancestral secreted mucins, the MUC1-N subunit contributes to a physical protective barrier. The adaptation of incorporating a non-mucin transmembrane M1C subunit afforded another distinct dimension of protection against inflammation [62]. Inclusion of the MUC1 SEA domain and the capacity for auto-proteolytic cleavage of the translated full-length MUC1 protein thus enabled this separation of functions. Of note, the MUC1-C subunit, designated here as M1C, is devoid of glycosylated tandem repeats and, by definition, is not a mucin (Fig. 3).

The MUC1 transmembrane and cytoplasmic domain sequences are highly conserved in eutherian mammals; whereas, significant differences can be found in the tandem repeats of the extracellular domain [35,36]. For example, the aa sequences and number of tandem repeats differ in elephant Muc1 (Loxodonta Africana) and Homo sapiens MUC1; whereas, the M1C region is highly conserved (NCBI gene database). These variations in extracellular domain sequences may be attributable as adaptations to differences in diet and environmental exposures rather than cancer propensity. In this regard and unrelated to Muc1 extracellular variations, elephants are apparently protected against developing cancer by amplification of the p53 tumor suppressor gene for promoting reproductive fitness [64].

3.2. Activation of the MUC1-N/M1C heterodimer at the epithelial apical cell membrane

How the MUC1-N/M1C complex is transported from the ER to the apical cell membrane of barrier epithelial cells has been of interest in terms of potentially targeting this pathway in malignant cells [65]. The anterior gradient-2 (AGR2) protein disulfide isomerase (PDI) processes and traffics MUC1 to the cell membrane [66]. MUC1 also traffics via apical recycling endosomes (AREs) to the apical surface [67]. At the apical cell membrane, the MUC1-N/M1C heterodimer acts as a sensor of entropic forces within the extracellular matrix. With loss of homeostasis, increases in mechanical forces disrupt the complex with (i) release of MUC1-N beyond the glycocalyx to form a protective physical mucous barrier, and (ii) activation of M1C to induce cellular responses, such as loss of polarity and lineage plasticity, to stress [62,68–70](Fig. 4).

Fig. 4.

Activation of the MUC1-N/M1C complex at the apical cell membrane. MUC1-N/M1C complexes are expressed at the apical borders of polarized barrier epithelial where they are poised to respond to biotic (i.e., viruses) and abiotic (i.e., environmental toxin) insults. MUC1-N extends beyond the glycocalyx while tethered to the transmembrane M1C subunit. With loss of homeostasis, MUC1-N is released from the complex into a physically protective mucous barrier that captures pathogens and toxins for ciliary clearance. Additionally, M1C is activated and forms homodimers to initiate repair by promoting disruption of the epithelial cell layer and induction of EMT. Modified from Kufe, 2009 [32].

3.3. M1C disrupts integrity of the epithelial cell layer in the response to stress

Epithelia resist mechanical stress to maintain tissue integrity that is disrupted by malignant transformation [71–73]. The structure and function of the mammalian epithelial cell layer is maintained by intercellular adhesion complexes that include adherens junctions (AJs), tight junctions, and desmosomes. The AJ comprised E-cadherin and cytoplasmic catenin cofactors (α, β, γ, and p120-catenin) is of importance for stabilizing cell-cell adhesion and conserving the integrity of epithelial tissues [71,74]. Activation of M1C suppresses E-cadherin expression [75,76] and forms complexes with β-, γ- and p120-catenin [77–83]. As a result, M1C functions in disrupting the AJ complex and thereby the epithelial cell layer in the response to stress, which is in principle reversible with resolution of the insult, but sustained in malignant transformation [62].

3.4. M1C promotes loss of apical-basal polarity

Positioning of M1C at the apical domains of epithelial cells also regulates apical-basal polarity that is disrupted in cancer cells [72,73]. Control of cell polarity relies on three main protein complexes that include the (i) PAR network, (ii) CRUMBS/PLS1/PATJ complex and (iii) SCRIB/LGL/DLG complex [84–86].

In disrupting apical-basal polarity, M1C binds directly to the inflammatory NF-κB p65 transcription factor [87]. M1C/NF-κB complexes regulate activation of NF-κB target genes, which include the ZEB1 tumor suppressor [88]. In turn, M1C forms direct complexes with ZEB1 that repress the CRB3 gene and expression of the Crumbs complex CRB3 protein [89]. M1C also downregulates the PATJ protein of the CRUMBS complex [89]. Additionally, M1C/ZEB1 complexes repress the miR-200 gene, which encodes a tumor suppressor that reverses EMT [90]. M1C → NF-κB signaling coordinates these responses with activation of the LIN28B pathway, which suppresses expression of the miRNA let-7 that, like miR-200, inhibits EMT [91,92]. M1C also drives EMT by inducing the EMT transcription factors (EMT-TFs) TWIST1 and SNAIL [93]. M1C thereby integrates disruption of the AJ junction with loss of apical-basal polarity and induction of EMT in malignant progression.

In summary, the MUC1-N/M1C heterodimer consisting of mucin and non-mucin subunits localizes to the apical membranes of polarized epithelial cells where it functions as a sensor of stress. Activation of M1C by loss of homeostasis induces a wound healing response by driving lineage plasticity that includes disruption of intercellular adhesion complexes and induction of EMT necessary for initiating repair of the insult. These responses are reversible with wound healing, whereas aberrant M1C expression in cancer cells drives irrevocable lineage plasticity that is a hallmark of malignant transformation.

4. M1C activates signaling at the cell membrane

Protecting eutherian mammals from biotic and abiotic insults was not an elementary undertaking that necessitated complex adaptations and exaptations. Among these, wound repair, which includes inflammatory, proliferative and remodeling phases, is critical for survival. M1C contributes to each of these phases by integrating signals at the cell membrane with intracellular pathways in the cytoplasm and nucleus that are necessary for repair. Despite this protective function, prolonged dysregulation of M1C by chronic inflammation promotes irreversible activation of these pathways in driving cancer progression [62]. The switch from a M1C protective function to one that confers malignant transformation is posited here to be the consequence of non-genetic alterations that become heritable with prolonged M1C activation by chronic inflammation. In this context, epigenetic determinants play causal roles in both cancer initiation and tumor progression [94–97].

4.1. M1C interacts with RTKs at the cell membrane

The M1C extracellular domain (M1C/ED) contains an Asn-36 NLT motif that is modified by N-glycosylation (Fig. 5) [98]. N-glycosylated M1C ranges in size from ~20 to 25 kDa as dictated by the extent of attached N-glycans; whereas, unglycosylated M1C is expressed as a 17 kDa protein [98](Fig. 5). N-glycosylation of the M1C NLT motif is of functional importance as a galectin binding site [98]. Galectins are a family of lectins containing conserved carbohydrate-recognition domains (CRDs) with specificity for β-galactosides on N- and O-linked glycans [99]. Among these, galectin-3 is a unique family member that is overexpressed in cancers and contributes to the formation of biomolecular condensates [99–101]. M1C increases galectin-3 expression by a post-transcriptional mechanism and, in turn, galectin-3 binds to M1C at the N-glycan-modified Asn-36 NLT site [98](Fig. 5). As a result, galectin-3 functions as a bridge in facilitating interaction of M1C and EGFR, as well as other RTKs, at the cell membrane (Fig. 5) [98].

Fig. 5.

Structural features of the M1C extracellular domain. M1C consists of a 58 aa extracellular domain (ED), a 28 aa transmembrane domain (TM) and a 72 aa cytoplasmic domain (CD). The M1C/ED includes a region that forms a non-covalent heterodimer with MUC1-N at the cell membrane. The M1C/ED also includes conserved alpha-3 and alpha-4 helices that are targets for recognition by aptamers, antibodies and small molecules. The NLT motif functions as a site for modification by N-glycosylation and thereby binding of galectin-3, which serves as a bridge for interactions with RTKs at the cell membrane, activation of intracellular effectors and the formation of biomolecular condensates. Modified from Kufe, 2022 [62].

As noted, under homeostatic conditions, M1C is expressed at the apical borders of polarized epithelial cells, whereas RTKs are positioned at the basolateral borders [32]. Disruption of homeostasis and thereby loss of apical-basal polarity facilitates interactions of M1C with EGFR and other RTKs that include ERBB2/HER2 [102], FGFR [103], MET [104–106], and insulin-like growth factor receptor-1 [32,62,72,107].

4.2. M1C functions as a scaffold for integrating proliferative and inflammatory signaling pathways

Studies of the M1C interaction with EGFR have provided comprehension of a functional role in the activation of RTK signaling (Fig. 6). EGFR is inhibited in its monomeric form; whereas, activation is achieved through formation of EGFR dimers and their cross-phosphorylation [108]. M1C homodimers form complexes with EGFR at the cell membrane that contribute to EGFR activation [109–112]. In turn, EGFR phosphorylates the M1C cytoplasmic domain on Y-46, which forms a binding site (pYEKV) for the SRC SH2 domain [109,113]. Phosphorylation of the M1C cytoplasmic domain on Y-8 by SRC forms a pYGQLD site for binding of the SHP2 phosphotyrosine phosphatase that enhances RTK-mediated RAS→ERK signaling [114]. In addition, phosphorylation of the M1C cytoplasmic domain on Y-20 (pYHPM) functions as a binding site for PI3K and activation of the AKT pathway (Fig. 6) [62,115].

Fig. 6.

Structural modifications of the M1C cytoplasmic domain drive signaling pathways in wound healing and malignant transformation. The M1C/CD is an intrinsically disordered 72 aa region as is often found in proteins that function as nodes in integrating signaling pathways. The CQC motif acts as a sensor of redox balance and is necessary for forming M1C homodimers, as well as M1C heterodimers with downstream effectors that include MYC, TCF4, TAK1 and JAK1. M1C/CD is modified by RTKs and TKs (ABL, SRC). As a result, the corresponding pY sites function as binding motifs for effectors with SH2 domains, such as SHP2, PI3K, SHC, PLCγ, SRC and GRB2, that contribute to RTK→RAS signaling. M1C/CD integrates the proliferative RTK→RAS pathway with effectors, such as NF-κB and STAT1/3 that drive inflammatory signaling in wound healing and malignant transformation. Modified from Kufe, 2022 [62].

Auto-phosphorylation of the EGFR cytoplasmic tail recruits adapter proteins, such as GRB2 [116]. EGFR-GRB2 complexes recruit SOS1 by binding of the GRB2 SH3 domains to the SOS1 proline-rich region. SOS1 then catalyzes nucleotide exchange of RAS-GDP to RAS-GTP [117]. Phosphorylation of the M1C cytoplasmic domain on Y-60 (pYTNP) functions as a binding motif for the GRB2 SH2 domain and the recruitment of SOS1 [118,119]. M1C also acts as a scaffold for binding of SHC and PLCγ that contribute to RTK and RAS signaling [62](Fig. 6). EGFR and GRB2 form a phase condensate that controls the ability of SOS1 to activate RAS [120,121]. Increasing evidence supports the involvement of M1C as a scaffold for the formation of biomolecular condensates that integrate RTK and RAS signaling in driving malignant transformation [62,73,100,120–123].

M1C/CD is also modified by the PKC and GSK3β serine-threonine kinases that together with EGFR and SRC regulate the downstream SAGNGGSSL motif which functions for the binding of β-catenin, NF-κB, and STAT1/3 (Fig. 6). In this way, the M1C/CD CQC motif and SAGNGGSSL region facilitate activation of the (i) TCF4 → β-catenin, (ii) TAK1 → NF-κB, and (iii) JAK → STAT1/3 signaling pathways. M1C activates proliferative and inflammatory pathways that contribute to wound healing. M1C/CD thus integrates (i) proliferative RTK → RAS and β-catenin signaling with (ii) inflammatory NF-κB and STAT1/3 TFs that are recognized drivers of malignant transformation. In this way, M1C has the capacity for promoting a state of chronic inflammation in association with the potential for heritable epigenetic alterations in cancer cells [124].

5. M1C transduces intracellular stress signaling

The involvement of M1C in activating proliferative and inflammatory signaling extends to downstream pathways in mitochondria and the nucleus. An important question for potentially targeting these pathways is how the transmembrane M1C protein enters the cytoplasm and is then transported intracellularly?

5.1. Internalization and recycling of M1C at the cell membrane

Internalization of MUC1 from the cell membrane is orchestrated by clathrin-mediated endocytosis [65,125–128]. The M1C cytoplasmic domain binds to adapter protein-2, which induces clathrin assembly and recruits endocytic accessory proteins [126,129]. As found for cell membrane-bound proteins released by depalmitoylation [130], modification of M1C by palmitoylation and binding of the adapter protein-1 regulates MUC1 recycling back to the cell membrane [131].

Cell membrane-bound proteins like M1C are recycled by Golgi-ER retrograde trafficking [130]. The M1C cytoplastic domain forms a direct complex with the disulfide-isomerase ER protein 5 (ERp5; PDIA6) [132]. Erp5 catalyzes disulfide bonds necessary for the formation of M1C homodimers [133]. M1C also forms a direct complex with the RAB27A GTPase that plays a role in docking and fusion of exosomes at the cell membrane [134–136]. The interaction between M1C and RAB27A contributes to secretion of M1C containing exosomes from cell membranes and thereby exosome-mediated extracellular functions [133,137]. These M1C interactions are therefore of significance in recycling M1C at the cell membrane and in the cytoplasm of cancer cells. An important distinction is that recycling of M1C is a separate process from that associated with disruption of MUC1-N/M1C complexes imbedded in the cell membrane. In this regard and in the absence of MUC1-N, M1C has the capacity to form structures in proximity to the cell membrane that function as scaffolds for integrating proliferative RTK/RAS signaling with activation of inflammatory pathways.

Recycling of M1C monomers at the cell membrane and in the cytoplasm would structurally expose the hydrophobic transmembrane domain. This potential limitation for intracellular signaling is circumvented by the formation of M1C homodimers and multimers. The M1C cytoplasmic domain CQC motif forms covalent Cys-mediated complexes between M1C monomers that stabilize higher order structures [138]. Formation of M1C homodimers is necessary for interacting with RTKs, transport by HSP70/90 to mitochondria and import by the nuclear pore complex (NPC) into the nucleoplasm [62,139–142](Fig. 7). M1C homodimers are thereby accessible for binding to molecular chaperones and delivery intracellularly to mitochondria and the nucleus.

Fig. 7.

M1C transduces proliferative and inflammatory signals from the cell membrane to mitochondria and the nucleus. Activation of M1C by loss of homeostasis drives the formation of M1C homodimers that are conferred by the M1C/CD CQC redox sensor and covalent disulfide bonds. M1C homodimers form complexes with RTKs at the cell membrane and function as a scaffold for activation of RTK→RAS signaling. Activated M1C homodimers undergo endocyclic recycling at the cell membrane and cytoplasm. In the cytoplasm, M1C homodimers are transported by HSP70/90 to mitochondria, where they promote survival and redox balance. M1C homodimers are imported into the nucleus by forming complexes with components, such as importin-beta, of the nuclear pore complex (NPC). In the nucleus, M1C localizes to nuclear speckles that include paraspeckles and PML bodies, which function in regulating gene expression. M1C is also embedded in chromatin as monomers and higher order structures. There, M1C interacts with TFs and effectors of epigenetic reprogramming in driving inflammatory, proliferative and remodeling responses that contribute to wound healing and, when irreversible, to malignant transformation. Modified from Kufe, 2009 [32].

5.2. M1C regulates mitochondrial functions in promoting redox balance and survival

M1C homodimers entering the cytoplasm by endocytic recycling are transported to mitochondria for regulating survival and redox balance [62,139,140,143–146](Fig. 7). M1C is transported to the mitochondrial outer membrane by HSP70/90 in the response to RTK activation and stress [103,139,140]. There, M1C blocks (i) release of mitochondrial apoptotic factors, (ii) activation of caspase-3 and (iii) induction of apoptosis [103,139,140]. M1C binds directly to the pro-apoptotic BAX protein and inhibits BAX function in inducing the release of cytochrome c and apoptosis [144]. M1C also interacts with the mitochondrial membrane ATPase family AAA domain-containing 3A (ATAD3A) protein, which is of importance for maintenance of mitochondrial DNA (mtDNA) structure and function [147,148]. M1C thereby promotes mitophagy by inducing degradation of ATAD3A and protecting PTEN-induced kinase 1 (Pink1) from ATAD3A-mediated cleavage [147].

5.3. M1C regulates oxidative phosphorylation

M1C also functions at the level of the inner mitochondrial membrane in regulating oxidative phosphorylation (OXPHOS) and redox balance critical for promoting stemness and the cancer stem cell (CSC) state. M1C regulates nuclear genes encoding effectors of mitochondrial Complexes I-V that are essential for OXPHOS [143]. M1C represses mtDNA genes encoding proteins in mitochondrial Complexes I-V [143]. In addition, M1C (i) downregulates the mitochondrial transcription factor A (TFAM) necessary for mtDNA gene transcription, and (ii) upregulates the mitochondrial transcription termination factor 3 (mTERF3) [143]. Increases in mtROS, largely in the form of superoxides occur by leakage of electrons from Complexes I, III and IV [149]. By repressing nuclear and mtDNA genes necessary for Complexes I, III and IV, M1C suppresses mitochondrial ROS production [143]. M1C also regulates ROS levels by integrating suppression of OXPHOS with activation of aerobic glycolysis [113,143,150,151]. Additionally, M1C plays an important role in linking hypoxia and metabolism with the regulation of oxidative phosphorylation [152].

M1C thus integrates RTK signaling at the cell membrane with activation of mitochondrial functions that contribute to survival by (i) attenuating stress-induced apoptosis and (ii) integrating energy production with redox balance. In this way, M1C confers proliferative and survival responses that have the adverse capacity to become irreversibly established in association with heritable epigenetic reprogramming of gene expression patterns.

6. M1C regulates gene expression in the nucleoplasm and chromatin

M1C homodimers are imported into the nucleus by the NPC [141,142](Fig. 7). In the nucleoplasm, M1C localizes to extrachromosomal paraspeckle condensates [142] (Fig. 7). M1C is also expressed in chromatin, where it interacts with TFs and effectors of epigenetic reprogramming [153,154](Fig. 7). M1C thereby has the capacity to drive gene expression patterns and epigenetic modifications that, when irreversible, contribute to malignant transformation.

6.1. M1C regulates gene expression by the formation of paraspeckles

As background for how M1C functions in regulating gene expression, paraspeckles are nuclear biomolecular condensates formed by the lncRNA NEAT1 and associated RNA binding proteins (RBPs) [155]. NEAT1 encodes (i) a 3.7 kb NEAT1_1 transcript that maintains genome integrity and is dispensable for paraspeckle formation, and (ii) a 23 kb NEAT1_2 isoform essential as a scaffold for binding of paraspeckle proteins [156–159]. NEAT1_2 recruits RBPs, including the splicing factor proline- and glutamine-rich (SFPQ), non-POU domain-containing octamer-binding protein (NONO and fused in sarcoma (FUS) among others essential for paraspeckle biogenesis, pre-mRNA splicing, DNA repair and transcriptional regulation [156,160–163]. Paraspeckles are increased in response to stress associated with inflammation, cell lineage transitions and cancer [162,164,165].

Like MUC1, paraspeckles are mammalian-specific structures [166]. The evolutionary concordance of M1C and paraspeckles in mammals is of interest in that M1C and NEAT1 are both activated by loss of homeostasis and both are overexpressed in human cancers [62,167]. The M1C → NF-κB and M1C → MYC pathways drive NEAT1_1 and NEAT1_2 expression [142]. Moreover, M1C → MYC signaling induces expression of SFPQ, NONO and FUS [142]. M1C accordingly functions in integrating regulation of NEAT1_2 and RBPs essential for paraspeckle formation. In addition, M1C forms complexes with SFPQ and FUS in the nucleoplasm and regulates their expression [142]. In accordance with these M1C dependencies, targeting M1C genetically and pharmacologically in cancer cells decreases paraspeckle formation [142].

M1C also localizes to PML bodies that, like paraspeckles, are nuclear biomolecular condensates restricted to mammals [168]. These findings have supported the need for additional studies to determine if M1C plays a role in alternative mRNA splicing in nuclear speckles as a mechanism of regulating gene expression.

6.2. M1C regulates gene expression in chromatin

M1C localizes to chromatin as homodimers and higher order multimers [153,154]. There, M1C interacts with TFs by forming direct complexes with TCF4/β-catenin [77,81,169], p53 [170], ZEB1 [88], TWIST1 [93], NF-κB [87], STAT1 [171], STAT3 [172], MYC [173] and hypoxia-inducible factor (HIF-1) [152]. M1C also interacts with multiple effectors of epigenetic reprogramming to regulate gene expression [62]. In cancer cells, the inflammatory M1C → NF-κB pathway regulates (i) DNMT1/3 and global DNA methylation [75], (ii) Polycomb Repressive Complex 1 (PRC1)/BMI1 which promotes ubiquitylation of H2A [174,175], (iii) PRC2/EZH2 that represses the expression of HOX and other genes by catalyzing H3K27 methylation [76,175,176], and (iv) the COMPASS family of H3K4 methyltransferases that counteract PRC1/2 [177–179](Fig. 8).

Fig. 8.

M1C integrates inflammatory and proliferative stress responses with chromatin remodeling in regulating gene expression and the CSC state. The inflammatory M1C → NF-κB pathway drives (i) DNMTs and global DNA methylation, (ii) PRC2/EZH2 and H3K27 methylation, and (iii) COMPASS and H3K4 methylation. COMPASS counteracts PRC2 regulation of gene expression. The proliferative M1C → E2F1 pathway activates the SWI/SNF BAF and PBAF chromatin remodeling complexes. M1C → E2F1 signaling also regulates EZH2, which is counteracted by the SWI/SNF complexes. M1C-induced activation of the inflammatory/NF-κB and proliferative/E2F1 pathways thus has the capacity to fine-tune the intersection of PRC2, COMPASS and SWI/SNF in regulating gene expression patterns. Of significance and in addition to driving transcription of PRC2 and SWI/SNF genes, M1C forms direct complexes with EZH2 and BAF/PBAF proteins and thereby regulates enhancer-like sequences (ELS) and expression of genes that contribute to stemness and the CSC state.

PRC1 and PRC2 play critical roles in cellular memory by repressing genes in stem cells and cancer [179–181]. Dysregulation of PRC1/2 also drives self-sustaining cell fate changes in cancer [179,181]. M1C integrates chronic inflammation with induction of PRC1/2 and their regulation by the COMPASS family [76,174,176,177]. M1C → E2F signaling regulates the SWI/SNF family of BAF and PBAF chromatin remodeling complexes that also oppose PRC1/2 by inducing an active chromatin state [178,179,182,183](Fig. 8).

As a result of these interactions with different effectors of epigenetic reprogramming, M1C regulates global changes in chromatin architecture across the genomes of cancer cells [184]. M1C induces differentially accessible regions (DARs) that associate with differentially expressed genes (DEGs) [184]. M1C-induced DARs align with gene enhancer regions regulated by (i) the JUN/AP-1 family of TFs, (ii) the BAF chromatin remodeling complex, and (iii) H3K27ac and H3K4me3 deposition [184]. M1C also forms direct complexes with JUN, which recruits SWI/SNF, in regulating chromatin accessibility and activation of enhancer-like sequences (ELS) (Fig. 8).

In summary and of relevance for promoting malignant transformation, M1C integrates stress signaling from the cell membrane with the regulation of gene expression in the nucleoplasm and chromatin. We posit that these epigenetic alterations become established by irreversible changes in heritable gene expression patterns that evolved for reproductive fitness and have been exploited by cancer cells. In this way, malignant transformation driven by M1C-induced epigenetic dysregulation is sustained by the inheritance of altered cancer cell fates [185–187].

7. M1C promotes pan-cancer progression

A central theme here is that M1C-driven responses to stress are, in principle, reversible with resolution of the insult. In the setting of chronic inflammation, prolonged and irreversible activation of M1C promotes epigenetic alterations that become established and irrevocable in association with driving malignant progression [62,73]. This concept as related to wound repair is not new in having been proposed initially by Rudolf Virchow in the 19th century [188] and later revisited by Harold Dvorak as “cancer is a wound that does not heal” [189].

7.1. M1C drives cancers of stem cell lineages in barrier tissues

Early work in the MUC1 field largely focused on the highly prevalent adenocarcinomas of the breast, lung, prostate and gastrointestinal tract [32,72,190,191]. The MUC1 gene is localized to 1q22, which is a region frequently amplified in human cancers [32,62,72]. Upregulation of MUC1 expression in these carcinomas associates with poor clinical outcomes [192]. Interestingly, mutations in the MUC1 gene, including the region encoding the cytoplasmic domain, are uncommon and none have been specifically linked to malignant progression. One potential exception is a T112P mutation in the MUC1 extracellular domain that has been identified in pancreatic and certain other adenocarcinomas, although the functional significance of this alteration is unclear [193].

Over time, a role for M1C was also uncovered in promoting adenocarcinomas with neuroendocrine (NE) dedifferentiation or transdifferentiation [191]. M1C dictates NE lineage specification in NE prostate cancer (NEPC) [194], pancreatic ductal carcinomas with NE features (PDAC-NE) [195], and rare pancreatic NE tumors (PNET) [196]. M1C also drives stemness in small cell lung cancer (SCLC) [197] and Merkel Cell Carcinoma (MCC) [198].

These findings extended to hematologic malignancies that, for example, include myeloid leukemias, multiple myeloma and lymphomas encompassing distinct hematopoietic lineages [199]. Involvement of M1C in hematologic malignancies was at first glance incongruent with the notion that MUC1 evolved to protect barrier epithelia. Upon reassessment of that reasoning, barrier tissues contain diverse resident stem cells of epithelial, as well as hematopoietic, origin that respond to loss of homeostasis and are thereby subject to chronic inflammation and cancer progression. These resident stem cells are also subject to clonal hematopoiesis of indeterminate potential (CHIP), which has been linked to hematologic malignancies and solid tumors [200,201].

A common thread throughout these malignancies is addiction to M1C for self-renewal capacity and tumorigenicity that extends to cancers associated with chronic viral infections.

7.2. Involvement of M1C in virus associated cancers

Infections with Epstein-Barr virus (EBV), human papillomavirus (HPV), hepatitis B virus (HBV), hepatitis C virus (HCV), Kaposi’s sarcoma-associated herpesvirus (KSHV), Merkel Cell polyomavirus (MCPyV) and human T-cell lymphotropic virus (HTLV-1) are associated with approximately 10% of cancers worldwide [202]. MUC1 evolved in mammals to protect against viruses; whereas, prolonged M1C activation in settings of chronic viral infections has the potential for promoting malignant transformation [62].

7.2.1. EBV

As one notable example, M1C is activated in EBV associated gastric cancer (EBVaGC) cells [203]. EBV was discovered in 1964 in cells cultured from a Burkitt’s lymphoma (BL) biopsy [204]. Approximately 250,000 cases of cancer and 2% of cancer deaths each year are attributed to EBV-associated malignancies [205]. Among these, EBV contributes to the pathogenesis of epithelial carcinomas that include about 10% of gastric carcinomas [206] and ~ 95% of undifferentiated nasopharyngeal carcinomas (NPCs) in the Far East [207].

EBNA1 is the only EBV protein uniformly expressed in EBV-associated cancers [208]. Despite decades of EBV research, it has remained unclear how EBNA1 contributes to cancer progression [208]. Intriguingly, EBNA1 forms an auto-regulatory complex with M1C that controls the expression of host cell and EBV latency genes [203]. In this way, EBNA1 exploits M1C to maintain EBV latency and, in turn, prolonged activation of M1C in response to chronic EBV infection promotes EBVaGC progression [203]. EBV positive and -negative GCs are dependent on M1C for self-renewal [203]. In addition, EBV positivity has no significant effect on survival of patients with GCs [203], invoking the possibility that, in contrast to M1C, dependence on EBNA1 may be lost after induction of the EBVaGC malignant phenotype. These findings provide initial evidence that M1C could contribute to other EBV associated epithelial cancers.

7.2.2. HPV

M1C plays a role in the progression of squamous cell carcinomas (SCCs) derived from stratified squamous epithelia [154,209]. The barrier function of squamous epithelia in the head and neck is disrupted by (i) exposure to cigarette smoke and alcohol, and (ii) HPV infections with progression to head and neck squamous cell carcinomas (HNSCCs) [209,210], supporting the importance of chronic inflammation in HNSCC progression. M1C is activated in HPV-positive HNSCC cells and is necessary for their survival [154]. MUC1 is also expressed in the malignant cell populations of HPV-positive HNSCC tumors [154]; however, it is not known if HPV proteins co-opt M1C for driving the HNSCC malignant phenotype. Interestingly, M1C is activated in HPV-negative HNSCC cells and is necessary for their self-renewal capacity [154]. Additionally, MUC1 is upregulated in primary and metastatic HPV-positive cervical tumors [211] in support of further study to determine if M1C contributes to self-renewal and tumorigenicity of these cancers.

7.2.3. MCPyV

M1C is broadly expressed in Merkel cell cancers (MCCs) and at higher levels in MCPyV-positive (MCCP) relative to MCPyV-negative (MCCN) tumors [198]. In MCCP and MCCN cell lines, M1C regulates common transcriptomes and is necessary for their self-renewal [198]. These findings that MCCP and MCCN cells are addicted to M1C provide the basis for determining if MCPyV drives chronic inflammation and thereby M1C activation in promoting the MCCP malignant phenotype [198].

The above available evidence, while limited to date, lends support to the further study of M1C in linking chronic inflammation with progression of virus-associated cancers.

7.3. Chronic activation of M1C by inflammation and replication stress drives cancer progression

A remarkably common finding in multiple cancers is that M1C forms self-perpetuating auto-inductive pathways in which M1C is activated by inflammation and, in turn, sustains a chronic inflammatory response [62,73]. M1C dictates chronic induction of the type I and II IFN pathways in triple-negative breast cancer (TNBC) cells in association with conferring DNA damage resistance and an immunosuppressive phenotype [190,212,213]. In castration-resistant prostate cancer (CRPC), M1C integrates activation the type II IFN pathway with chromatin remodeling and immunosuppression [182,183,191]. M1C also activates the type I and II IFN pathways in (i) tyrosine kinase inhibitor (TKI) resistant non-small cell lung cancer (NSCLC) cells [153], (ii) pancreatic ductal carcinoma cells (PDAC) cells [195], and (iii) HPV-positive and HPV-negative HNSCCs [154]. M1C-dependent activation of the type I and II IFN pathways has been linked to genomic instability and induction of STING [214]. In turn, M1C → STING signaling induces type I interferons, STAT1 and IFN stimulated genes (ISGs) that confer DNA damage resistance and immune evasion [213,214].

Dependence on M1C in certain cancers is associated with DNA replicative stress and genomic instability [62,73]. M1C activates the E2F and MYC proliferative pathways in driving DNA replicative stress in (i) PDAC [195], (ii) SCLC [197], (iii) MCC [198], (iv) colorectal cancer (CRC) [114,215], and (v) PNET cells [196]. The M1C-driven inflammatory and proliferative pathways are not necessarily mutually exclusive [195] and both tend to be activated in M1C-dependent cancers. M1C-driven inflammatory and proliferative signaling in cancers likely represents their dysregulation in the wound healing response [62,73]. Along these lines, M1C integrates activation of inflammatory and proliferative pathways with changes in chromatin architecture and the epigenome in association with driving the CSC state [62,73,176,177,184].

How M1C drives pan-cancers is thus supported by involvement in activating inflammatory, proliferative, and remodeling responses [62,73]. These responses encompass M1C coordinated activation and epigenetic reprogramming of diverse signaling pathways that, if prolonged and irreversible, become heritable in promoting malignant transformation.

8. M1C integrates pan-cancer progression with treatment resistance

The role of M1C in barrier tissues extends to protecting cancer cells from insults conferred by (i) treatment with cytotoxic and targeted agents, and (ii) recognition by the immune system. M1C is necessary for the CSC state in pan-cancers. The CSC state confers resistance to treatment and immune evasion [216–219]. In this sense, the M1C-driven CSC state has been appropriated to confer resistance to treatment with cytotoxic, targeted and immunotherapeutic agents. A critical question is how M1C confers resistance to agents with diverse structures and pleotropic mechanisms of anti-cancer activity. An answer relates to a fundamental M1C capacity for conferring heritable epigenetic modifications that recall and thereby protect against insults imposed on cancer cells [187,220,221].

8.1. M1C confers resistance to genotoxic anti-cancer agents

Early studies identified a role for MUC1 in conferring resistance of cancer cells to cisplatin and etoposide, which was attributed to suppression of the DNA damage-induced intrinsic mitochondrial pathway [139]. These findings were extended to doxorubicin, paclitaxel and gemcitabine by the demonstration that targeting MUC1/M1C reverses resistance to these agents [93,222–224]. Involvement of M1C in conferring resistance to cytotoxic agents with diverse structures and mechanisms of action was also attributed to promoting DNA damage repair, upregulation of ABC drug transporters and the induction of aerobic metabolism [223–226]. A role in promoting DNA damage repair was further supported by the demonstration that MUC1/M1C confers resistance to ionizing radiation [227–229]. Suppression of the response to genotoxic anti-cancer agents aligned with the association of MUC1 expression and poor clinical outcomes in patients with pan-cancers [192,230–237].

8.2. M1C promotes resistance to targeted anti-cancer agents

Early work identified involvement of M1C in conferring resistance to (i) tamoxifen in ER+ breast cancer [238], (ii) trastuzumab in HER2+ breast cancer [102], (iii) afatinib in EGFR mutant NSCLC [111], (iv) bortezomib in multiple myeloma [239], and (v) midostaurin in FLT3 mutant acute myeloid leukemia cells [240]. Subsequent studies uncovered a role for M1C in resistance of (i) BRAF(V600E) mutant colorectal cancer (CRC) cells to BRAF inhibitors [114], (ii) EGFR mutant NSCLC cells to the third-generation EGFR inhibitor osimertinib [153], (iii) CRPC cells to enzalutamide [151], (iv) HR+/HER2- BC cells to endocrine therapy (ET) and the CDK4/6 inhibitor abemaciclib [241], and (v) PDAC KRAS G12D mutant cells to KRAS inhibitors [242]. These findings of resistance to a diversity of cytotoxic and targeted anti-cancer agents invoked the notion that M1C activates networks broadly wired for protecting against therapeutic challenges. This capacity extends to agents that induce ER stress and anoikis [243,244] and may be integrated with the highly glycosylated MUC1 extracellular domain that can modulate the permeation of drugs across the cell membrane [245].

8.3. M1C integrates the CSC state with inflammatory memory

Resident stem cells and CSCs have the capacity for inflammatory memory; that is, the recall of previous insults which effectively informs responses to subsequent exposures [60]. Inflammatory memory of CSCs is clearly a major challenge that can confer resistance to further treatment, albeit by unclear mechanisms.

Drug treatment persister (DTP) cells arise from a population with characteristics of CSCs that serve as founders for disease relapse [246,247]. Reversibility of DTPs has implicated epigenetic mechanisms as being responsible for their lineage plasticity and survival [246,247]. A provocative discovery was that NSCLC mutant EGFR DTPs resistant to TKIs by pleotropic genetic mechanisms are dependent on M1C for survival [153]. M1C confers resistance of NSCLC TKI DTPs to EGFR mutant targeted agents by regulating STAT1 and the IFN type I/II pathways [248].

The establishment and recall of NSCLC TKI resistance is conferred by activation of MUC1 at proximal enhancer-like sequences (pELSs) [248]. MUC1 pELS-1 is activated by M1C and STAT1; whereas, pELS-2 is activated by M1C, JUN/AP-1 and PBAF occupancy. The MUC1 pELS-1/2 regions are further activated by increases in (i) chromatin accessibility, (ii) PBAF/PBRM1-mediated chromatin remodeling, and (iii) H3K27ac and H3K4me1 deposition [248]. Noteworthy is that these M1C-induced epigenetic alterations align with those identified in the activation of inflammatory and proliferative pathways (Fig. 8). The MUC1 pELS regions in turn function as memory domains for driving STAT1 and downstream IFN-stimulated genes (ISGs) that confer TKI resistance [248].

How M1C functions as a common effector of resistance to cytotoxic, targeted and immunotherapeutic agents can thus be explained, at least in part, by induction of an inflammatory memory response to treatment-associated stress. This M1C function could extend beyond CSCs to the activation macrophages and other immune effectors [249]. Another fundamental component of M1C-activated inflammatory memory could reside in the induction of retrotransposons and A3s that promote genomic instability and mutagenesis [246,250–253]. In this regard, an active area of investigation is based on the discoveries that the M1C/STAT1 pathway drives retrotransposon and A3 expression in association with resistance [253].

9. Development of anti-MUC1 therapeutics for cancer treatment

The development of agents targeting MUC1 has been a challenge. Early work focused on MUC1-N as a target based on uniqueness of the variable numbers of 20 amino acid tandem repeats. O-glycosylation of the repeats is often modified by dysregulation of glycosyltransferase expression in cancer cells [254,255]. As a result, the MUC1-N tandem repeats can be decorated with incomplete or truncated glycans that represent potential tumor-associated antigens [254]. Incomplete O-glycosylation can also result in the exposure of cryptic peptide epitopes in the repeats [256]. Nonetheless, heterogeneity of O-glycosylation patterns and shedding of MUC1-N into the circulation have complicated targeting of the MUC1-N tandem repeats. As an alternative MUC1 target, M1C is not modified by O-glycosylation and is not shed from the cancer cell surface. M1C also has druggable extracellular and cytoplasmic domains.

9.1. Targeting the MUC1-N subunit

9.1.1. Anti-MUC1-N CAR T cells

One concept of targeting tumor-associated MUC1-N was proposed using modified T cells [257,258]. CAR T cells were engineered against the tumor-associated Tn-glycoform expressed on the MUC1-N tandem repeats by cancer cells [259,260]. Based on highly effective anti-tumor activity in preclinical models [259], this approach was advanced to clinical evaluation by Tmunity/Kite in patients with advanced cancers; but, was closed because of dose-limiting skin toxicity (Table 1).

Table 1.

MUC1 Therapeutic Approaches.

	Sponsor	Phase	Status
MUC1-N
CAR T Cells	TMUNITY/KITE	I	Closed
	CART-TnMUC1	Solid Tumors	DLT:SKIN
	NCT04025216
ADC	DAIICHI SANKYO	I/II	Active
	DS-3939	Solid Tumors
	NCT05875168
VACCINE	MERCK KGaA	II	Active
	TECEMOTIDE	HER2-BC
	NCT01507103
M1C
CAR T Cells	POSEIDA/ROCHE	I/II	Active
	P-MUC1C-ALLO1	Solid Tumors	Not Recruiting
	NCT05239143
	MINERVA	I/II	Active
	huMNC2	Breast Cancer	Last Update 6/18/2023
	NCT04020575

Notably, MUC1-N targeted CAR T cells were found to be effective and well-tolerated in mouse tumor models [259,261]. These discrepancies in pre-clinical and clinical toxicity of MUC1-N targeted CAR T cells may reside in the partial 34% homology of the Muc1-N/MUC1-N tandem repeat sequences in mice and humans [262]. Accordingly, the human MUC1 transgenic mouse represents a potentially more informative model for pre-clinical toxicology assessments of anti-MUC1-N agents. MUC1-N levels in the plasma of patients with breast and other cancers are often increased as assessed by the CA-15-3 assay [263,264]. As a result, circulating levels of MUC1-N with aberrant tandem repeat O-glycosylation in patients represents a potential barrier to targeting MUC1-N on the surface of cancer cells.

9.1.2. Anti-MUC1-N ADCs

The issue of overcoming circulating MUC1-N pools also pertains to other approaches, such as ADCs, that target the MUC1-N tandem repeats. Daiichi Sankyo DS-3939 is an ADC developed from the Glycotope antibody directed against the TA-MUC1 antigen expressed on MUC1-N tandem repeats [265]. DS-3939 is conjugated to the exatecan derivative DXd via tetrapeptide-based cleavable linkers. DS-3939 is being evaluated in a Phase I/II clinical trial for patients with advanced MUC1-N-expressing cancers (Table 1) [265]. Initial findings have confirmed 10 partial responses in patients with NSCLC, ovarian and breast cancers without significant issues of safety and tolerability (NCT05875168).

9.1.3. Anti-MUC1-N vaccines

The MUC1-N tandem repeats have been the target of the clinically evaluated tecemotide, TG4010 and PANVAC vaccines.

9.1.3.1 Tecemotide (L-BLP25, Stimuvax) is a liposomal-based vaccine incorporating a MUC1 tandem repeat peptide synthesized to induce responses against HLA class I and II epitopes [266]. The Phase III START trial of tecemotide as maintenance therapy in patients with stage III NSCLC demonstrated improvement in OS versus placebo (30.8 vs 20.8 months) in a subgroup of patients who had received concurrent chemoradiotherapy [267]. A Phase II trial of tecemotide and bevacizumab for patients with stage III NSCLC treated with chemoradiation demonstrated tolerability of the combination and encouraging improvements in PFS and OS [268].

A subsequent report in patients with HER2- early breast cancer demonstrated that addition of tecemotide to neoadjuvant systemic therapy in a two-arm ABCSG34 randomized trial markedly improves distant recurrence free survival (DRFS) and overall survival (OS) (Table 1) [269,270]. Notably, this is the first report of a survival benefit for a vaccine in breast cancer patients, supporting the performance of additional studies to extend these findings and define the underlying mechanisms.

9.1.3.2 TG4010 is a modified vaccinia Ankara (MVA) vaccine expressing MUC1-N and interleukin-2 [271]. Administration of TG4010 in combination with chemotherapy has improved clinical outcomes as compared to chemotherapy alone in patients with NSCLC [272,273]. In the TIME randomized Phase IIb trial assessing efficacy of TG4010 with chemotherapy in patients with NSCLC, TG4010 significantly improved survival as compared to placebo in association with inducing anti-MUC1 T cell responses [274–276].

9.1.3.3 PANVAC is a vaccine combining MVA and fowlpox viral vectors co-expressing MUC1 and CEA with the TRI-COM B7–1, ICAM-1 and LFA-3 costimulatory molecules [277]. A novel PANVAC-VF prime and boost strategy evaluated in clinical trials for patients with advanced carcinomas demonstrated induction of immunologic responses. In a Phase II randomized clinical trial for patients with metastatic breast cancer, combining PANVAC-VF with docetaxel increased PFS (7.9 months) vs docetaxel alone (3.9 months)(HR = 0.65) [278]. Patients in both arms developed measurable immunity to MUC1 and CEA [278].

9.2. Targeting the M1C subunit

The M1C subunit includes a 58 aa extracellular domain (ED) that represents a potential target for the generation of monoclonal antibodies (Fig. 5). In contrast to MUC1-N that is released from the cancer cell membrane, M1C is not shed in association with undetectable levels in the circulation. The discovery that M1C/ED has two conserved alpha helices which become exposed with activation provided the basis for generating the 3D1 monoclonal antibody against the alpha-3 helix (Fig. 5) [279].

9.2.1. Anti-M1C CAR T cells

Sequences derived from the humanized hu3D1 antibody have been used for development of a fully allogeneic CAR-T cell therapy, designated P-MUC1C-ALLO1 [280]. P-MUC1C-ALLO1 was designed to enrich for CAR T cells with stem cell memory that are self-renewing, multipotent and long-lived. Administration of P-MUC1C-ALLO1 resulted in complete regressions of human TNBC and ovarian tumor xenografts in mouse models [280]. Based on these preclinical data, P-MUC1C-ALLO1 was evaluated in a Phase I, dose-escalation trial in patients with advanced or metastatic solid tumors (NCT05239143). P-MUC1C-ALLO1 has been well tolerated in initial cohorts, which is of interest in contrast to the anti-MUC1-N CAR T cell trial that was closed because of dose-limiting skin toxicity (Table 1). However, like other allogeneic CAR T cell approaches for the treatment of solid tumors, administration of P-MUC1C-ALLO1 to date has been associated with limited clinical activity [280].

9.2.2. Anti-M1C ADCs

As another approach for targeting M1C, the humanized hu3D1 antibody was conjugated to MMAE using a cleavable linker to generate an anti-M1C ADC [279]. The M1C ADC exhibited anti-tumor activity in (i) human MUC1 transgenic mice harboring syngeneic MUC1-expressing tumors, (ii) nude mice bearing human ZR-75–1 HR+/HER2- BC tumor xenografts, and (iii) NCG mice engrafted with a human TNBC PDX model. These findings in the absence of significant associated toxicities provided support for further development of the M1C ADC. In accomplishing this, work on the M1C ADC was advanced by the NCI Cancer Moonshot Immuno-Oncology Translational Network (IOTN) to enable translation from preclinical development to clinical evaluation [281]. In addition, the NCI NExT Program assumed responsibility for performing the M1C ADC IND-enabling studies. Based on activity of the M1C ADC in human TNBC [279] and HR+/HER2- [241] PDX models, a Phase I dose-escalation trial of the M1C ADC trial is being sponsored by the NCI CTEP Program for the treatment of patients with refractory HER2- BCs. Support by the NCI IOTN, NExT and CTEP has thus collectively made possible overcoming the “gap” between preclinical and clinical development of the M1C ADC. The findings that the M1C ADC is also highly effective against treatment-resistant NSCLC, CRPC/NEPC and PDAC PDX models has provided further support for clinical evaluation of the M1C ADC in patients with these refractory cancers [151,242,248].

9.2.3. Anti-M1C vaccines

Unlike MUC1-N as a potential target for an anti-MUC1 vaccine as described above, efforts in developing a vaccine against M1C have been limited. M1C contains 7 potential CD8+ CTL epitopes that span HLA-A2, HLA-A3, and HLA-A24 MHC class I alleles [282]. Enhancer agonist peptides for each of these epitopes have been identified for incorporation into vaccine platforms and the ex vivo generation of anti-M1C T cells [282]. Liposome-based mRNA vaccines targeting M1C represent an additional approach for generating CTLs against these epitopes.

9.2.4. M1C small molecule inhibitors

The M1C cytoplasmic domain (CD) is a 72 aa intrinsically disordered protein [283]. M1C/CD includes a CQC motif that is necessary for the formation of M1C homodimers and higher order multimers [62,73]. The GO-203 cell-penetrating peptide was developed to block the CQC motif and thereby M1C function. GO-203 has proven to be a highly informative tool compound for validating the CQC motif as a druggable target in multiple in vitro and in vivo studies [62,73]. M1C expression has been targeted with the translation inhibitor silvestrol [284], anti-sense oligonucleotides [198] and the CSC inhibitor salinomycin [285]. However, attempts at directly targeting the M1C CQC motif or other regions in the cytoplasmic domain with small molecule inhibitors have been unsuccessful to date [286], emphasizing the importance of efforts to address this unmet need. Of translational importance, a recurrent finding throughout M1C studies of human cancer cells has been that targeting M1C reverses the epigenetic changes that confer malignant transformation and treatment resistance.

MUC1 was identified over 40 years ago. Advances in the field have identified the MUC1-N and M1C subunits as potential targets for the treatment of MUC1-addicted cancers. Much of the work on developing anti-MUC1 agents has focused on targeting MUC1-N with vaccines and, more recently, CAR T cells and an ADC. MUC1-N presents certain obstacles, specifically, shedding of this subunit from the cancer cell surface and the circulating pools that need to be overcome to target cancer cells. Progress is being made in developing agents against the non-shed M1C subunit with the demonstration that the extracellular and cytoplasmic domains are validated targets for cancer treatment.

10. Conclusions and future perspectives

The evolution of MUC1 in eutherian mammals contributes to reproductive fitness as evidenced by roles in placentation and in lactation. MUC1 also contributes to survival of the newborn and beyond to reproductive age by protecting barrier tissues from biotic and abiotic insults. In this sense, the appearance of MUC1 in mammals was arguably an important event in the subsequent evolution of primates. Nonetheless, whether MUC1 contributed to the hallmarks of primates, such as enlarged brains, is not known.

The evidence that MUC1 increases fitness during reproductive age, while paradoxically contributing to cancer is consistent with the concept of antagonistic pleiotropy [287,288]. The contribution of a gene, like MUC1, to reproductive success combined with promoting cancer later in life could be considered of importance to the increasing rates of cancer in older individuals [25]; but not necessarily the increasing incidence of early onset cancers [15]. Rather, early onset cancers are more likely the result of evolutionary mismatches; a setting in which environmental changes differ from those to which a species has adapted, resulting in increased stress and susceptibility to cancer [16,17].

Chronic activation of M1C by biotic and abiotic insults could be a contributing factor to the increasing number of new cancer cases in both younger and older individuals [11]. Cancer incidence, as defined by number of newly diagnosed cases per capita per year, is declining for certain cancers, whereas others, such as colorectal cancers, are increasing in younger age groups [14,289,290]. Strikingly, in people younger than 50 years in the US, mortality has increased for those with CRCs [291]. Insults to barrier epithelia are associated with exposures to tobacco smoke, alcohol consumption and dietary risk factors, including ultraprocessed foods that have been linked to chronic inflammation [11,15,292–295]. These factors are largely behavioral and can be reduced by changes in lifestyle. By contrast, exposure of barrier tissues to environmental insults, such as (i) air pollutants [296], microplastics [297], and perfluorochemicals [298], among others, that constitute the “exposome”, has been largely unavoidable [299].

In summary, this review posits that M1C contributes to the increasing incidence of human cancers by antagonistic pleiotropy and/or evolutionary mismatches in association with the highly prevalent emergence of chronic inflammation. Dating back to the work of Rudolf Virchow, chronic inflammation has been increasingly associated with cancers in humans [300,301]. M1C evolved in mammals to protect against inflammation. In humans, prolonged activation of M1C in settings of chronic inflammation results in epigenetic alterations that become heritable in driving malignant transformation and therapeutic resistance.

Data availability

Data will be made available on request.