Stem Cell Competition

Abstract

Cell competition is a conserved, fitness-based process in which “winner” cells eliminate or displace less-fit “loser” neighbors, thereby shaping development, tissue homeostasis, and disease. This review synthesizes how principles of competition operate in stem cell compartments, focusing on the adult Drosophila testis as a genetically tractable model where neutral competition provides a baseline for stem cell turnover, and defined genetic perturbations bias competition to favor specific clones. I compare intra- and inter-lineage competition among somatic cyst stem cells, germline stem cells, and dedifferentiating germ cells, highlighting mechanisms that link signaling, proliferation, adhesion, and niche remodeling to competitive fitness. I then discuss mammalian paradigms such as clonal hematopoiesis, intestinal “supercompetitors,” and spermatogonial stem cell dynamics in open niches, framing stem cell competition as a driver of aging, clonal evolution, and cancer risk.

Introduction

Competition between neighboring cells has been documented in organisms from Drosophila to humans, where it shapes a wide range of biological processes, from removing suboptimal cells during development to maintaining tissue homeostasis in adulthood (reviewed in (Claveria & Torres, 2016; Khandekar & Ellis, 2024; Morata & Calleja, 2020; Vishwakarma & Piddini, 2020)). Mosaic analysis of developing Drosophila epithelia first revealed this phenomenon: cell heterozygous for a ribosomal protein mutation (Rp/+) were viable in a uniform population, but they were eliminated when surrounded by wild-type (+/+) neighbors (Morata & Ripoll, 1975). In this context, the Rp/+ cells became “losers,” and the wild-type cells that outcompeted them were termed “winners”. This fitness-dependent removal of otherwise viable cells came to be defined as “cell competition”, now referred to as “active competition” (Bowling, Lawlor, & Rodriguez, 2019; van Neerven & Vermeulen, 2023). Work in developing Drosophila epithelia also revealed the process of “supercompetition”, in which wild-type cells are outcompeted by fitter, oncogene-expressing cells (de la Cova, Abril, Bellosta, Gallant, & Johnston, 2004; Moreno & Basler, 2004). In supercompetition, the wild-type cells are the losers, and the oncogene-expressing cells are the winners.

Cell competition is conserved in mammals. In mosaic mouse embryos with cells expressing different levels of the transcription factor Myc, cells with lower Myc levels were eliminated (Claveria, Giovinazzo, Sierra, & Torres, 2013). Similarly, embryonic stem cells that were tetraploid or that were deficient in Bone Morphogenetic Protein (BMP) signaling or in autophagy are eliminated by wild-type cells (Sancho et al., 2013).

Over the past two decades, researchers have investigated a different type of competition, namely stem cell competition, in adult tissues and its role in homeostasis. Historically, stem cells were generally viewed as slow-cycling and label-retaining (Cotsarelis, Sun, & Lavker, 1990). However, more recently, it has been appreciated that numerous tissues are sustained by cycling stem cells that may reside in a defined niche, like intestinal stem cells (ISCs), or in a specific location in the tissue, like epidermal stem cells, or in an open niche, like spermatogonial stem cells (SSCs) in the mammalian testis (Ellis et al., 2019; Kitadate et al., 2019; Snippert et al., 2010).

To maintain long-term homeostasis in renewing tissues, stem cells use one of two self-renewal strategies: asymmetric division or population asymmetry (Watt & Hogan, 2000). In asymmetric division, every stem cell division consistently produces two daughters with different fates, one remains a stem cell and the other differentiates (Gonczy, 2008). In population asymmetry, the balance between self-renewal and differentiation is regulated across the entire stem cell pool: some stem cells differentiate and are lost, while others divide to replace them. Under population asymmetry, homeostasis is maintained when these events occur at equal frequencies, keeping the overall number of stem cells stable even though individual stem cells may not persist indefinitely (reviewed in (Simons & Clevers, 2011)).

To distinguish between these self-renewal strategies, the stem cell and its descendants must be labeled genetically using fluorescent markers, and these clones must be analyzed for size and stability over time. If a tissue is maintained by asymmetrically dividing stem cells, then there should be a stable population of clones that are long-lived and that are of a typical size. However, if a tissue is maintained by population asymmetry, then there should be fewer clones over time, and these clones should be increasing in size (Klein & Simons, 2011). Analysis of multiple tissues in mice and flies, including the intestine and testis, revealed that they are maintained by population asymmetric self-renewal involving an equipotent stem cell population with stochastic fate, in which mitosis can give rise to two stem cells, one stem cell and one differentiating cell, or two differentiating cells (Amoyel, Simons, & Bach, 2014; Clayton et al., 2007; de Navascues et al., 2012; Klein, Nakagawa, Ichikawa, Yoshida, & Simons, 2010; Lopez-Garcia, Klein, Simons, & Winton, 2010). This clone behavior was termed “neutral competition” and refers to the stochastic loss and replacement of cells, driven by chance cellular processes, that together maintain a stable number of cells within a tissue.

Lineage-tracing analyses and modeling revealed that “biased competition” arises when stem cells do not have equal probabilities of self-renewal and loss, causing certain clones to expand at the expense of others. Unlike neutral competition, where stochastic loss and replacement events keep all stem cells equally fit, non-neutral competition reflects intrinsic or acquired differences between cells, such as mutations, altered signaling responsiveness, metabolic changes, or microenvironmental advantages (Amoyel et al., 2014; de Navascues et al., 2012; Snippert, Schepers, van Es, Simons, & Clevers, 2014; Vermeulen et al., 2013). These disparities shift the balance of outcomes so that “advantaged” stem cells are more likely to persist and proliferate, while neighboring cells are more likely to differentiate or be lost from the stem cell pool. As a result, specific clones can dominate the niche over time, reshaping tissue composition and potentially contributing to aging, disease, or tumor initiation.

The Adult Drosophila testis

The adult Drosophila melanogaster testis provides an elegant and tractable system for studying how stem cell niches maintain tissue homeostasis and how intrinsic and extrinsic factors govern stem cell behavior. The adult testis is a coiled tube approximately 2 mm in length and 0.1 mm in diameter. It is capped at the apical end and connects basally with other components of the reproductive tract, with progressive stages of germ cell development distributed along the apical-basal axis (Fuller, 1993; Hardy, Tokuyasu, Lindsley, & Garavito, 1979). At the apical tip of the testis, a well-defined niche microenvironment composed of hub cells, germline stem cells (GSCs), and somatic cyst stem cells (CySCs) orchestrate continuous sperm production throughout adult life (Fig. 1). These populations form a discrete and spatially stable signaling environment that integrates JAK-STAT, Hedgehog (Hh), BMP, adhesion, and transcriptional cues to regulate self-renewal, differentiation, and competition between stem cells.

Figure 1: Schematic of the adult Drosophila testis.

The testis is enclosed by a smooth muscle sheath and an underlying basal lamina (gray). Approximately 12 hub cells (yellow) form the stem cell niche, some of which are anchored to the basal lamina of the overlying muscle sheath. Germline stem cells (GSCs; magenta) contact hub cells and divide with oriented mitosis to produce a self-renewing GSC and a differentiating gonialblast (Gb). The Gb undergoes transit-amplifying divisions to generate 16-cell spermatogonial cysts, which then enter meiosis and differentiate into sperm. Spermatogonial stages (light purple) can be distinguished by the morphology of the germ cell–specific organelle, the fusome: dot-shaped in GSCs and Gbs, dumbbell-shaped in 2-cell spermatogonia, and branched in 4-, 8-, and 16-cell spermatogonia. Cyst stem cells (CySCs; dark blue) contact hub cells via thin cytoplasmic extensions and divide to produce early cyst cells (light blue) that exit the niche. Two early cyst cells encapsulate each Gb, and as the enclosed germ cells undergo transit-amplifying divisions, the surrounding somatic cells differentiate into late cyst cells (light blue cells that are further from the hub).

Cell types in the adult Drosophila testis

Hub Cells

Hub cells comprise a small, non-dividing cluster at the testis apex and serve as the anatomical and functional anchor of the niche (Fig. 1, yellow cells) (Hardy et al., 1979). They are attached to the basal lamina of the muscle sheath of the tissue via integrin dependent adhesion (S. Lee, Zhou, Kim, Kalbfleisch, & Schock, 2008; Tanentzapf, Devenport, Godt, & Brown, 2007). Hub cells secrete multiple short-range self-renewal cues to maintain two unipotent stem cell populations, GSCs that give rise to spermatids and CySCs that produce somatic support cells termed cyst cells. Hub-produced self-renewal cues include Decapentaplegic (Dpp) and Glass bottom boat (Gbb), BMPs that are bona fide self-renewal cues for GSCs (Fig. 2) (Kawase, Wong, Ding, & Xie, 2004; Shivdasani & Ingham, 2003). Hub cells also produce the cytokine Unpaired (Upd), which activates JAK-STAT signaling in GSCs and CySCs (Fig. 2) (Kiger, Jones, Schulz, Rogers, & Fuller, 2001; Tulina & Matunis, 2001). STAT activation is not required for self-renewal in GSCs per se but instead mediates E-Cadherin (E-Cad)-dependent adhesion to hub cells (Leatherman & Dinardo, 2008, 2010). By contrast, STAT activation is essential for the self-renewal of CySCs (Issigonis et al., 2009; Kiger et al., 2001; Leatherman & Dinardo, 2008). Hub cells also produce Hh, which acts on CySCs but not GSCs to promote their self-renewal (Fig. 2) (Amoyel, Sanny, Burel, & Bach, 2013; Forbes, Lin, Ingham, & Spradling, 1996; Michel, Kupinski, Raabe, & Bokel, 2012).

Figure 2: Signaling pathways in the adult Drosophila testis.

Hub cells (yellow) produce Bone morphogenetic proteins (BMPs), the JAK/STAT ligand Unpaired (Upd), and the Hedgehog (Hh) pathway ligand Hh.

GSCs (pink) transduce BMP signaling through phosphorylation of the transcription factor Mad. BMP signaling represses expression of the germline differentiation factor bam, thereby promoting GSC self-renewal (circular arrow). BMP signaling also represses expression of the transcriptional co-repressor Hairless (H). H in turn represses expression of Imp, an RNA-binding protein. Imp promotes expression of Chinmo. High levels of Chinmo repress GSC competition, while loss of Chinmo promotes GSC competition. The lipase Brummer has also be shown to repress GSC competition.

CySCs (blue) activate transcription factors STAT and Ci, effectors of the JAK/STAT and Hh pathways. Activated phospho-STAT (pSTAT) activates target genes including Socs36E, Chinmo, and Zfh1; Ci also activates Zfh1. STAT through Chinmo and Zfh1 promotes CySC self-renewal (circular arrow). Zfh1 represses expression of Hippo pathway components Salvador (Sav) and Kibra (not shown), which restricts Yorkie (Yki) activation to Zfh1-positive cells. Yki promotes CySC self-renewal. Additionally, activation of Yki induces CySC-CySC competition. Socs36E is an induced negative feedback inhibitor of the JAK/STAT pathway, and it also inhibits Egfr/Ras/MAPK signaling. Egfr signaling in CySCs is activated by the EGF ligand Spitz (Spi) from germ cells. Egfr/Ras/MAPK is required for CySC self-renewal but plays a stronger role in CySC-CySC competition. Abl represses CySC-CySC.

The relative sizes of the activating and inhibitory arrows in CySCs indicate whether the factor more strongly regulates self-renewal or intra-lineage competition.

Adhesion molecules such as Ecad maintain the physical association of stem cells with the hub as GSCs or CySCs deficient for Ecad are not maintained in the niche microenvironment (Voog, D’Alterio, & Jones, 2008; Yamashita, Jones, & Fuller, 2003). Because hub cells are quiescent, their fixed number sets an upper limit on the number of stem cells that can compete for contact, making them a central determinant of niche capacity and competitive dynamics.

Germline Stem Cells

Approximately 6–12 GSCs form a rosette around the hub (Fig. 1, magenta cells in contact with hub cells) (Yamashita et al., 2003). They express the germline marker Vasa. They undergo oriented division in which one daughter remains attached to the hub and retains stem cell identity, while the distal daughter becomes a gonialblast (Gb) and initiates differentiation (Yamashita et al., 2003). GSCs asymmetrically inherit centrosomes, with the mother centrosome remaining with the GSC while the daughter centrosome is inherited by the Gb (Yamashita, Mahowald, Perlin, & Fuller, 2007). There have been reports of asymmetrical inheritance of histones between the GSC and the Gb (V. Tran, Lim, Xie, & Chen, 2012; Xie et al., 2015), but this has recently been disputed (Li, Tong, Ohlstein, & Guo, 2025). GSC identity requires continuous exposure to Dpp and Gbb produced by both hub cells and CySCs; BMP signaling in GSCs represses the differentiation factor bag of marbles (bam) and maintains stemness (Kawase et al., 2004; Leatherman & Dinardo, 2008, 2010; Shivdasani & Ingham, 2003) (Fig. 2). Loss of BMP signaling results in premature GSC differentiation (Kawase et al., 2004; Shivdasani & Ingham, 2003). BMP signaling has recently been reported to repress expression of the transcriptional co-repressor Hairless (H) in GSCs. In turn, H represses transcription of Imp, encoding an RNA-binding protein that promotes expression of the transcription factor Chinmo. This BMP-H-Imp-Chinmo axis regulates GSC-GSC competition (Fig. 2) (Zheng et al., 2025). Additionally, the triglyceride lipase Brummer also been reported to control GSC competitiveness (Fig. 2) (Chao et al., 2022).

CySCs and GSCs do not have robust synchronization of their cell cycles (Ben-Hamo-Arad et al., 2026; Lenhart & DiNardo, 2015), but the production of their differentiated progeny is tightly coordinated. Each differentiating GSC daughter (i.e., the Gb), must be paired with precisely two somatic cyst cells generated from the CySC pool, creating a 2:1 soma-to-germline ratio that is indispensable for germ cell differentiation (Kiger, White-Cooper, & Fuller, 2000; J. Tran, Brenner, & DiNardo, 2000). Following somatic encapsulation, the Gb undergoes four rounds of transit-amplifying divisions with incomplete cytokinesis, yielding 2-, 4-, 8-, and ultimately 16-cell spermatogonial cysts (Fuller, 1993). Disruption of these soma-germline interactions compromises germline differentiation, and the loss of even a single cyst cell from a developing cyst is sufficient to block spermatogenesis (Fairchild, Islam, & Tanentzapf, 2017; Sarkar et al., 2007; Smendziuk, Messenberg, Vogl, & Tanentzapf, 2015). Because cyst cells cease dividing after differentiation and instead enclose the germline through extensive cytoskeletal and membrane remodeling, the requisite 2:1 ratio is thought to be established within the niche before the Gb moves away from the CySC population that provides its somatic partners.

The stage of germ line differentiation can be identified the shape of a germ cell specific organelle - the fusome - that is dot-shaped in GSCs and Gbs, dumbbell-shaped in 2-cell spermatogonia and is branched in 4-, 8- or 16-cell spermatogonia (Fig. 1, light purple cells) (Hime, Brill, & Fuller, 1996). The 16-cell spermatogonia then transitions into a pre-meiotic spermatocyte that undergoes a protracted 3-day S phase, followed by two meiotic divisions. The resulting 64 haploid germ cells then enter spermiogenesis, a process during which nearly all organelles are restructured, and the flagellar tail is assembled, giving rise to mature spermatids (Fabian & Brill, 2012; Fuller, 1993; White-Cooper, 2010).

The germline in the Drosophila testis exhibits a remarkable degree of plasticity. Spermatogonia can dedifferentiate into GSCs under homeostatic conditions and during regeneration after tissue loss (Brawley & Matunis, 2004; Cheng et al., 2008; Herrera & Bach, 2018; Sheng, Brawley, & Matunis, 2009). This plasticity is likely results from the similarity in gene expression among early germ cells as two independent single-cell transcriptomic studies could not distinguish GSCs from early spermatogonia on the basis of differential mRNA expression (Raz et al., 2023; Witt, Benjamin, Svetec, & Zhao, 2019).

Cyst Stem Cells

There are ~36–44 somatic cells that express the stem cell marker Zfh1 and lack the differentiation marker Eyes absent (Eya) (Amoyel et al., 2013; Inaba, Yuan, & Yamashita, 2011; Leatherman & Dinardo, 2008). Zfh1-positive cells form two rows surrounding the niche in the adult testis. Of these, approximately 13 are somatic CySCs, defined as Zfh1-positive cells making physical contact with the niche (Fig. 1, dark blue cells with thin membranes contacting hub cells) (Amoyel et al., 2014). CySCs act as an extended niche for GSCs by secreting Dpp and Gbb (Leatherman & Dinardo, 2010). CySCs are the only mitotic somatic cells in the testis; they divide and self-renew in response to Hh and Upd produced by hub cells (Amoyel et al., 2013; Gonczy & DiNardo, 1996; Hardy et al., 1979; Kiger et al., 2001; Leatherman & Dinardo, 2008; Michel et al., 2012; Tulina & Matunis, 2001). CySCs express the JAK/STAT target Chinmo and the JAK/STAT and Hh target Zfh1; both Chinmo and Zfh1 are required for CySC niche residence (Fig. 2) (Albert, Puretskaia, Terekhanova, Labudina, & Bokel, 2018; Flaherty et al., 2010; Leatherman & Dinardo, 2008). CySCs are part of the extended niche for GSCs as a result of their secretion of BMPs downstream of STAT through Chinmo and Zfh1 (Flaherty et al., 2010; Leatherman & Dinardo, 2010). Hyperactivation of the JAK/STAT pathway or overexpression of Chinmo or of Zfh1 in CySCs causes an expansion of the CySC population, and the expanded CySC population maintains GSC-like cells away from the niche (Flaherty et al., 2010; Leatherman & Dinardo, 2008, 2010)

Chinmo promotes male somatic sexual identity in CySCs by repressing activation of the female somatic sex-determination pathway (Grmai, Hudry, Miguel-Aliaga, & Bach, 2018; Harsh, Liu, Bhaskar, Rushlow, & Bach, 2025; Ma, Wawersik, & Matunis, 2014). Zfh1 binds and downregulates salvador and kibra, encoding two tumor suppressor genes in the Hippo (Hpo) pathway, a conserved regulator of cell proliferation (Fig. 2) (Albert et al., 2018; Pan, 2010). As a result, Zfh1 restricts Yki activation and proliferation to Zfh1-positive somatic stem cells. Consistently, the Hpo pathway effector Yorkie (Yki) is required for CySC niche residency (Amoyel et al., 2014). Germ cells produce the EGF-like ligand Spitz (Spi). Spi activates Egfr/Ras/Mitogen-Activated Protein Kinase (MAPK) signaling in CySCs, and activation of this pathway is also required in CySCs for niche residency (Fig. 2) (Amoyel, Anderson, Suisse, Glasner, & Bach, 2016; Kiger et al., 2001; Singh, Liu, Zhao, Zeng, & Hou, 2016; J. Tran et al., 2000). As we will see below, Hpo, Hh, Egfr/Ras/MAPK pathways play important roles in CySC-CySC competition.

CySCs give rise to cyst cells that exit the cell cycle, encapsulate differentiating germ cells, and provide essential cues for their differentiation. As cyst cells differentiate, they downregulate Zfh1 and upregulate Eya (Fabrizio, Boyle, & DiNardo, 2003; Leatherman & Dinardo, 2008; Sarkar et al., 2007). Egfr signaling is required in cyst cells for proper differentiation, for encystment of the developing germline, and for promotion of GSC-gonialblast abscission (Kiger et al., 2001; Lenhart & DiNardo, 2015; Sarkar et al., 2007; Schulz et al., 2004; J. Tran et al., 2000). As germ cells progress to the 4-cell stage, the surrounding somatic cyst cells undergo a key maturation step, forming septate junctions with one another that generate a permeability barrier isolating the developing germline syncytium from the external milieu (Fairchild et al., 2017; Fairchild, Smendziuk, & Tanentzapf, 2015). Beyond serving as a physical and signaling scaffold, cyst cells also play an active metabolic role: they compartmentalize systemic carbohydrate metabolism and transfer critical metabolites, such as lactate, to adjacent germ cells. This somatic metabolic support is essential for germ cell survival and sustained spermatogenesis (de la Maza, Jefferson, Brucker, Paoli, & Amoyel, 2025).

In contrast to the early germline lineage, CySCs can be transcriptionally distinguished from their differentiating cyst cell daughters (Raz et al., 2023). Recent work has shown that CySC proliferation (cell-cycle progression) and differentiation (exit from stemness) are coupled through a metabolic checkpoint: the cell cycle-promoting E2f1/Dp transcription factor must be silenced to permit differentiation, because persistent E2f1/Dp activity suppresses mitochondrial gene expression and thereby interferes with differentiation, even when cells have exited the cell cycle (Sainz de la Maza, Hof-Michel, Phillimore, Bokel, & Amoyel, 2022). These findings could be interpreted to suggest that the CySC lineage is less plastic than the GSC lineage. However, lineage-tracing and functional studies have shown that somatic cells in the first and second rows can exchange positions, indicating that they form a pool of stem cells with at least some degree of plasticity (Chiang, Yang, & Yamashita, 2017; Yuen, Hillion, Wang, & Amoyel, 2021). Indeed, neighbor associations and cellular metabolic state can prime CySCs for differentiation (Amoyel, Hillion, Margolis, & Bach, 2016; Wang, Roiuk, Storer, Teleman, & Amoyel, 2025).

Types of competition in the Drosophila testis

This review will discuss four types of competition in the Drosophila testis: (1) intra-lineage competition between CySCs (i.e., CySC-CySC); (2) inter-lineage competition between CySCs and GSCs (i.e., CySC-GSC); (3) intra-lineage competition between GSCs (i.e., GSC-GSC; and (4) inter-lineage competition between dedifferentiating germ cells and resident stem cells.

Clonal mosaic analyses in the Drosophila testis

Lineage tracing is a powerful experimental approach used to label the descendants of a single cell, which enables mapping the relationships of cells within a tissue or organism (Kretzschmar & Watt, 2012). By permanently marking a specific cell or group of cells, typically through genetic labeling, inducible recombination systems, or live-imaging reporters, researchers can follow their progeny over time and determine how these descendants contribute to tissue growth, maintenance, and/or repair. As we will see below, lineage tracing has revealed that many tissues once thought to rely on stereotyped, asymmetric divisions instead employ population-level strategies such as neutral competition, where equipotent stem cells stochastically self-renew or differentiate. Lineage tracing also enables the reconstruction of differentiation hierarchies, and insights into how mutations or aging alter stem-cell behavior. Ultimately, lineage-tracing studies provide a dynamic and quantitative view of cell behavior in vivo, illuminating principles of development, homeostasis, and disease.

In Drosophila, lineage tracing has been greatly facilitated by sophisticated genetic tools that enable precise labeling and manipulation of small, well-defined cell populations. The FLP/FRT system induces mitotic recombination at engineered FRT sites following expression of the FLP recombinase, typically triggered by heat shock, generating genetically distinct daughter cells that can be selectively marked with reporters such as green fluorescent protein (GFP) (Xu & Rubin, 1993). The Mosaic Analysis with a Repressible Cell Marker (MARCM) technique allows positive labeling of homozygous mutant clones in an otherwise heterozygous background (T. Lee & Luo, 2001). In the germline, GSC clones can also be generated using a nanos (nos)-based flip-out strategy to drive gene depletion or overexpression in positively labeled GSCs (X. Ma et al., 2014). For all of these approaches, appropriate control clones are essential: controls must carry the same FRT sites for FLP/FRT and MARCM experiments, or express neutral proteins or control RNAi transgenes in the case of nos flip-out clones. Together, these tools have transformed the study of stem cells during tissue homeostasis in Drosophila. They enable visualization of mutant phenotypes at single-cell resolution, dissection of cell-autonomous versus non-autonomous mechanisms, and quantitative analysis of clonal behavior over time.

CySC-CySC competition in the Drosophila testis

Characteristics of neutrally competing stem cells

Sparse labeling of stem-cell clones is essential for accurately measuring clonal dynamics because it allows each marked lineage to be tracked independently, without interference from neighboring labeled cells. When too many stem cell clones are induced simultaneously, clones overlap, making it impossible to determine whether changes in clone size reflect intrinsic stem-cell behaviors, competitive interactions, or the simple merging of adjacent lineages. By inducing recombination in only a small fraction of stem cells, individual clones can be followed over time, enabling precise quantification of whether they persist or are lost and how often labeled stem cells replace one another. This sparse-labeling strategy is foundational to the neutral competition framework established in epithelial and germline tissues in mammals (Clayton et al., 2007; Klein et al., 2010; Lopez-Garcia et al., 2010) and is not merely a technical preference but a conceptual requirement for extracting the statistical principles that govern stem-cell behavior.

In their classic review, Klein and Simons outline several core assumptions that define how a homeostatic stem-cell population behaves (Klein & Simons, 2011). First is equipotency: all stem cells within the compartment are functionally equivalent, with no intrinsic hierarchy or stable fate bias. Each cell has the same average probability of self-renewal or differentiation, and no long-lived “elite” subset exists. Second is stochastic fate decisions. Each division produces one of three outcomes at random: symmetric renewal (two stem cells), symmetric differentiation (two daughters that leave the niche), or an asymmetric outcome (one stem cell daughter and one daughter that leaves the niche) that leaves stem-cell number unchanged (Fig. 3A). These stochastic events allow individual stem cells to be transient even as the population persists. Third is constant population size: homeostasis is maintained because losses of individual stem cells are balanced by duplication of neighboring ones. Thus, the population size is preserved not through invariant asymmetric division but through population-level balancing of gain and loss. Fourth is neutral competition: clones expand or are lost purely by chance, without selective advantages among stem cells under steady-state conditions. Over long timescales, clones drift toward extinction, or toward dominance that can lead to monoclonality. Fifth is predictable clonal dynamics: lineage-tracing data conform to universal statistical signatures. The mean clone size increases linearly with time, the fraction of surviving clones declines inversely with time, and clone size distributions collapse onto a common scaling curve when time is normalized by tissue-specific parameters such as division or loss rates.

Figure 3. CySC division outcomes and neutral competition in the adult Drosophila testis.

(A) A marked CySC (gray) undergoes division, with possible outcomes shown at right. In asymmetric renewal (top), the two daughters generate one self-renewing CySC (gray) and one differentiating cyst cell (light gray), which ensheaths a gonialblast together with an unmarked cyst cell (light blue). In duplication (middle), both marked daughters (gray) remain at the niche as CySCs, a process that requires displacement of a resident CySC (not depicted). In differentiation (bottom), both daughters of the marked CySC differentiate into cyst cells (light gray).

(B) Neutral competition. Under steady-state conditions, a labeled CySC clone (gray) may expand or be lost purely by chance. Most clones are lost over time, whereas rare persisting clones can expand, ultimately leading to monoclonality of the stem cell pool.

(C) Confocal image of a testis with CySC clone generated by the MARCM technique and labeled with membrane-targeted CD8-GFP (green) and analyzed at 2 days post-clone induction (dpci). (C′) Higher magnification showing Zfh1-positive CySCs (red), some of which contact the hub (DE-cadherin, blue) via thin cytoplasmic extensions (arrow).

(D-F) Confocal images of testes with control MARCM CySC clones expressing CD8-GFP (green) analyzed at 2 (D) and 14 (E, F) dpci. Clones are small at 2 dpci but exhibit substantial size variability by 14 dpci. Vasa (red) labels germ cells, Zfh1 (blue) marks CySCs and early cyst cells, and the hub is outlined by a dotted line.

(G) Comparison of experimental data and model predictions. The graph shows the mean fraction of labeled CySCs in persisting clones over time (boxes). The solid black line represents the best fit of the neutral competition model, while the dashed orange line shows the predicted clonal evolution following induction of a single CySC clone with a 3-day time shift using the same parameters. Clone size distributions from multiple independent induction events and from a single induction event converge rapidly. Error bars denote SEM.

(H) Distribution of persisting control MARCM clone sizes in wild-type testes at 2, 7, 14, and 28 dpci. Boxes show experimental data, and lines indicate model predictions. Error bars denote SEM.

C-H were reproduced and used with permission from the European Molecular Biology Organization from “Neutral competition of stem cells is skewed by proliferative changes downstream of Hh and Hpo”, by M. Amoyel, B. D. Simons, and E. A. Bach, 2014. The EMBO journal 33, 2295–2313; permission conveyed through Copyright Clearance Center, Inc.

CySCs conform to neutral competition dynamics

CySC are defined as Zfh1-positive cells contacting hub cells (Fig. 3C, C’, arrow). To determine whether CySCs in the Drosophila testis follow a model of invariant asymmetry or population asymmetry, sparse wild-type clones expressing membrane-bound GFP were generated (Amoyel et al., 2014). Analysis at two days post clone induction (dpci) confirmed a low induction frequency, a prerequisite for accurate modeling (Fig. 3D, D’). The total numbers of GFP-labeled and unlabeled Zfh1-positive cells in contact with the hub were quantified, and the percentage of labeled CySCs per testis was calculated. Testes were then classified as “persisting”, if at least one GFP-positive CySC maintained hub contact, or “differentiating”, if all labeled cells had detached from the hub.

If CySCs divide by invariant asymmetry, the mean fraction of labeled CySCs within persisting clones should remain constant over time. In contrast, under a population asymmetry (i.e., neutral competition) model, the fraction of labeled cells should progressively increase as unlabeled CySCs are lost. Consistent with the latter, persisting clones showed a time-dependent increase in the fraction of labeled CySCs (Fig. 3G); this is exemplified by the markedly varied fraction of labeled cells shown in Figs. 3E, E’, F, F’ at 14 dpci (Amoyel et al., 2014). This result is inconsistent with invariant asymmetry but is consistent with stochastic loss and replacement dynamics.

To quantify these dynamics, a mathematical framework previously developed for mammalian intestinal stem cells was applied (Lopez-Garcia et al., 2010). The model assumes: (1) the CySC pool consists of equipotent stem cells with equal probabilities of loss and replacement; (2) CySCs form a one-dimensional “necklace” around the hub; and (3) total stem cell number remains constant, such that each loss through differentiation is exactly balanced by duplication of a neighboring CySC. By fitting the experimental data (Fig. 3G, boxes) to model predictions (Fig. 3G, black line), the CySC loss/replacement rate was estimated at ~1 per day (Amoyel et al., 2014). Using this rate, the model accurately recapitulated the cumulative clone size distribution observed in wild-type testes (Fig. 3H) (Amoyel et al., 2014). Together, these findings establish that ~13 CySCs per testis form an equipotent population maintained by neutral competition (Fig. 3B), mirroring the stochastic dynamics of mammalian stem cells.

Mutations that accelerate proliferation bias stem cell competition

Building on this neutral baseline, the next question was whether specific mutations could bias competition to favor the mutant stem cell. The Hh pathway provided a paradigm for testing this possibility. Hh signaling is required for CySCs to remain in the niche, and lineage-wide loss of pathway activity causes a reduction but not total loss of the CySC pool (Amoyel et al., 2013; Michel et al., 2012). To determine whether excessive Hh activity confers a competitive advantage, clones mutant for patched (ptc), which encodes the Hh receptor, were analyzed. Ptc-mutant cells experience constitutive, ligand-independent activation of the pathway (Y. Chen & Struhl, 1996). Sparse induction of ptc-mutant clones (Fig. 4A) revealed that by 14 dpci, ptc-mutant clones contained more labeled CySCs on average and frequently occupied the entire somatic lineage (Fig. 4B) (Amoyel et al., 2014). ptc-mutant CySCs displaced neighboring wild-type cells, effectively “winning” the competition for niche occupancy as testes containing ptc-mutant clones exhibited significantly fewer unlabeled (i.e., wild-type) CySCs. Notably, ptc-mutant CySCs retained the ability to differentiate into cyst cells that ensheath spermatogonia (Michel et al., 2012), indicating that these mutants are not locked in a self-renewing state. Thus, ptc-mutant CySCs possess a competitive advantage over wild-type cells during stem cell replacement, progressively displacing them from the niche.

Figure 4. Biased competition of ptc-mutant CySCs in the adult Drosophila testis.

(A, B) Confocal images of testes with ptc-mutant MARCM clones expressing CD8-GFP (green) analyzed at 2 (A) and 14 dpci (B). Single-channel images show the clone (A′, B′), germ cells labeled with Vasa (red; A″, B″), and CySCs and early cyst cells labeled with Zfh1 (blue; A‴, B‴). The hub is indicated by a dotted line. Arrow (B-B‴) highlights displacement of GSCs by ptc-mutant CySCs.

(C) Time-dependent variation in the average size of ptc-mutant CySC clones. Data points (boxes) represent the mean fraction of labeled CySCs in persisting clones. The solid black line shows a fit of a neutral competition model modified to include a competitive bias favoring the labeled cells. The dashed orange line indicates the predicted clonal evolution following induction of a single CySC clone with a 3-day time shift using the same parameters. Clone size distributions from multiple independent induction events and from a single induction event converge rapidly. Error bars denote SEM.

(D) Distribution of persistent ptc-mutant clone sizes. Boxes show experimental data, and lines indicate model predictions. Error bars denote SEM.

(E) Biased competition model. A CySC clone with enhanced competitive fitness (green) is more likely to displace neighboring stem cells, resulting in increased fixation and the emergence of monoclonal stem cell pools. Such fitness advantages arise from elevated Hh, Yki, or MAPK activity, loss of Socs36E or Abl, or enhanced cell-cycle progression driven by CycE and Stg.

A-D were reproduced and used with permission from the European Molecular Biology Organization from “Neutral competition of stem cells is skewed by proliferative changes downstream of Hh and Hpo, by M. Amoyel, B. D. Simons, and E. A. Bach, 2014. The EMBO journal 33, 2295–2313; permission conveyed through Copyright Clearance Center, Inc.

Modeling reveals biased competition

Mathematical modeling was used to determine whether the observed ptc-mutant behavior reflects a bias in the otherwise neutral competition process. The model assumed two key conditions: (1) when a CySC - either wild-type or ptc mutant - is lost to differentiation, a neighboring ptc-mutant CySC has a higher probability of replacing it; and (2) this advantage is maintained after each displacement event. Holding the wild-type CySC loss/replacement rate constant, the model was adjusted to include a ~35% bias favoring ptc-mutant replacement of neighbors. Simulated dynamics under these parameters showed excellent agreement with experimental data (Fig. 4C). Likewise, the cumulative clone size distributions predicted by the model closely matched those observed experimentally across all time points (Fig. 4D). These analyses confirm that ptc-mutant clone behavior reflects a biased competition model, in which enhanced persistence arises from a probabilistic shift favoring self-renewal of the mutant clone (Fig. 4E) (Amoyel et al., 2014).

Proliferation as a driver of competitive dominance

The dominance of ptc-mutant clones suggested that accelerated proliferation might underlie their competitive advantage. Indeed, ptc-mutant CySCs displayed higher S- and M-phase indices than controls, indicating an overall increase in proliferation rate rather than a shortening of a single cell-cycle phase (Amoyel et al., 2014; Michel et al., 2012). The competitive advantage of ptc-mutant CySCs depended on this proliferative capacity: mutant clone expansion was significantly reduced in backgrounds heterozygous for the cdc25 homolog string (stg) or for other cell-cycle regulators such as E2f and cdk2 (Amoyel et al., 2014). By contrast, the hypercompetitiveness of ptc-mutant CySC clones could not be suppressed by reducing dose of α Catenin, which connects E-Cad to the cytoskeleton (Sarpal et al., 2012), suggesting that adhesion is not a critical factor in CySC-CySC competition (Amoyel et al., 2014).

To test whether enhanced proliferation alone is sufficient to drive competition, clones overexpressing both the G1/S-phase factor Cyclin E (CycE) and the G2/M-phase factor Stg were generated. In developing epithelia, co-overexpression of CycE and Stg accelerates the cell cycle and increases cell number (Neufeld, de la Cruz, Johnston, & Edgar, 1998). Similarly, CySC clones overexpressing CycE and Stg expanded rapidly and displaced wild-type CySCs from the niche, phenocopying ptc-mutant clones (Amoyel et al., 2014). These findings demonstrate that proliferation downstream of ptc is both necessary and sufficient to induce competitive dominance. Altering the rate of cell division thus skews the stochastic process of loss and replacement in favor of faster-dividing CySCs.

Further confirmation of the model that biased competition is driven by the rate of proliferation was shown by making clones lacking the kinase Hpo, a conserved tumor-suppressor that controls cell proliferation. hpo-mutant CySC exhibited pronounced overproliferation and progressively colonized the niche at the expense of wild-type CySCs (Amoyel et al., 2014). In fact, within the margin of error, the dynamics of ptc-mutant clones and hpo-mutant clones were indistinguishable. Despite the expansion of hpo-mutant CySCs, hpo mutant cyst cells differentiated normally and surrounded spermatogonial cysts, like ptc-mutant cyst cells (Amoyel et al., 2014). Together, these findings demonstrate that increased proliferative activity, whether caused by hyperactivation of Hh signaling or by loss of Hpo, can bias the normally neutral competition of CySCs, leading to the clonal dominance of faster-dividing cells.

Other factors regulating CySC-CySC competition

Socs36E and Egfr/Ras/MAPK:

Socs36E, a Suppressor of Cytokine Signaling (SOCS) protein, is an induced negative feedback inhibitor of the JAK/STAT pathway. SOCS proteins harbor SH2- and E3-ligase domains, which bind to phosphorylated (i.e., activated) signal transduction components and then target them for degradation by Ubiquitin-dependent proteolysis (Alexander & Hilton, 2004). In addition to the JAK/STAT pathway, SOCS proteins can also inhibit other tyrosine kinase signaling pathways, including Egfr/Ras/MAPK (Callus & Mathey-Prevot, 2002; Herranz, Hong, Hung, Voorhoeve, & Cohen, 2012; Rawlings, Rennebeck, Harrison, Xi, & Harrison, 2004). One group reported that CySC clones lacking Socs36E occupy an increased proportion of the CySC pool, indicating increased competitiveness (Issigonis et al., 2009). Another group found that CySCs mutant for either of two alleles of Socs36E (Socs36E^EY and Socs36E^PZ) were recovered in higher numbers than controls at all time points, indicating that they had better self-renewal than controls and lower probability of being lost through neutral competition (Amoyel, Anderson, et al., 2016). Socs36E-mutant clones had entirely replaced all wild-type CySCs at the niche by 14 dpci, and approximately 80% of clones were fixed (or monoclonal), indicating that they had colonized the entire niche (Amoyel, Anderson, et al., 2016).

What accounts for the enhanced competitiveness of Socs36E-mutant CySCs? One study attributed this phenotype to elevated JAK-STAT signaling, which was proposed to increase integrin-dependent adhesion of mutant CySCs to the niche (Issigonis et al., 2009). However, subsequent work showed that clonal elevation of JAK-STAT activity or of adhesion alone is not sufficient to drive CySC-CySC competition (Amoyel et al., 2014). Consistent with a model in which proliferation rate is a key determinant of competitive fitness, reducing the dosage of the cell-cycle regulator stg significantly suppressed the competitive advantage of Socs36E mutant clones, as evidenced by a marked decrease in the frequency of fixed niches (Amoyel, Anderson, et al., 2016).

If increased JAK/STAT activity cannot account for competitiveness of Socs36E-mutant CySCs, what does? MAPK signaling was significantly elevated in Socs36E-mutant clones compared with neighboring wild-type CySCs (Amoyel, Anderson, et al., 2016). Importantly, this enhanced MAPK activity underlies their competitive advantage. CySC clones mutant for Sos (Son of sevenless), which encodes a Ras guanine nucleotide exchange factor required for MAPK activation, were not recovered at 7 dpci. However, when these clones were made doubly mutant for Socs36E and Sos, they persisted for at least two weeks, indicating long-term stem cell function (Amoyel, Anderson, et al., 2016). Despite this persistence, Socs36E, Sos double-mutant clones failed to outcompete neighboring CySCs: few mutant cells were present at the niche at 14 dpci, and no clones became fixed, indicating an inability to colonize the niche (Amoyel, Anderson, et al., 2016).

Together, these observations indicate that elevating MAPK signaling within individual CySCs enhances their competitive fitness. In line with this idea, CySC clones expressing constitutively active MAPK or Ras^V12 progressively displaced most wild-type CySCs by 14 dpci (Amoyel, Anderson, et al., 2016). Accordingly, these data identify MAPK signaling as a key determinant of CySC access to the niche and demonstrate that the competitive dominance of Socs36E-deficient CySCs is largely driven by elevated MAPK activity (Amoyel, Anderson, et al., 2016).

Abl tyrosine kinase (Abl):

CySC clones lacking Abl, a non-receptor tyrosine kinase that phosphorylates cell adhesion and cytoskeletal proteins, were more competitive than control controls because Abl-mutant CySC clones were recovered at significantly high rates at 10 dpci than age-matched control clones (Stine, Greenspan, Ramachandran, & Matunis, 2014). By 8 dpci, there were significantly more Abl-mutant CySCs compared to wild-type CySCs per testis, indicating that Abl-mutant CySCs were outcompeting CySCs from the niche (Stine et al., 2014). The ability of Abl-mutant CySCs to cause the displacement of wild-type CySCs was diminished by knocking down Ecad in the clone (Stine et al., 2014). Because these authors found that Abl-mutant CySCs have similar mitotic indices as control CySCs, it was concluded that adhesion and not proliferation was important to the competitive properties of CySCs lacking Abl (Stine et al., 2014). Whether individual niches can become clonally fixed by Abl-mutant cells was not reported.

Mlf1-adaptor molecule (Madm):

CySCs clones lacking Madm, a tumor suppressor, were recovered at significantly higher rates that control clones and outcompeted a larger portion of the CySC pool, suggesting increased proliferation and possibly competitive advantage to the clones (Singh et al., 2016). The authors attributed this competitive advantage to autonomous increases in Egfr signaling and integrin expression. However, since these factors were not tested in Madm-mutant clones, their formal role in Madm-dependent CySC-CySC competition remains unresolved.

In sum, increased proliferation appears to be the main driving force underlying CySC-CySC competition (Fig. 4E). This mechanism is shared by multiple factors that cause competition between CySCs, including gain of Hh, Yki or Egfr/Ras/MAPK, and accelerating the rate of proliferation in the absence of any mutation is sufficient to induce CySC-CySC competition. Clonal increases in adhesion are not sufficient to cause competition between CySCs, and in case where increased adhesion has been implicated, more detailed examination is needed.

Self-renewal versus competitive fitness

A fundamental question in stem cell biology is how intrinsic self-renewal capacity (i.e., the ability to divide and generate stem cell progeny) relates to competitive fitness (i.e., the ability to retain niche occupancy). Importantly, self-renewal is not purely cell autonomous but is strongly influenced by interactions with neighboring cells and access to local, niche-derived signals (Simons & Clevers, 2011). Stem cells that fail to compete effectively for niche access are displaced, lose exposure to short-range maintenance cues, and consequently differentiate. Conversely, stem cells with enhanced competitive fitness can progressively replace their neighbors, enabling their progeny to expand and, in some cases, dominate the entire niche.

Distinguishing whether a signaling pathway primarily regulates self-renewal or competitive fitness requires complementary experimental approaches. Lineage-wide loss-of-function experiments test whether a pathway is essential for maintaining the stem cell pool, whereas clonal gain-of-function assays assess whether increased pathway activity confers a relative advantage within a mixed population. When depletion of a pathway reduces, but does not abolish, the stem cell compartment, and clonal activation promotes competitive dominance, the pathway is perhaps best interpreted as a regulator of competition rather than an absolute requirement for self-renewal.

EGFR/Ras/MAPK signaling exemplifies this principle. Although CySC clones with elevated MAPK activity outcompete neighboring cells, uniform suppression of MAPK signaling across the lineage does not eliminate the entire CySC pool, even though MAPK-deficient clones are lost (Amoyel, Anderson, et al., 2016; Kiger et al., 2001; Singh et al., 2016; J. Tran et al., 2000). These observations argue against a strict requirement for MAPK signaling in CySC self-renewal and instead support a primary role in regulating competitive fitness. In contrast, JAK/STAT signaling displays the opposite behavior: clonal activation does not confer a competitive advantage, whereas lineage-wide inhibition leads to profound depletion of the CySC population, indicating an essential role in sustaining intrinsic self-renewal capacity (Amoyel et al., 2014; Brawley & Matunis, 2004; Kiger et al., 2001; Leatherman & Dinardo, 2008, 2010; Yuen et al., 2021).

Hh signaling behaves similarly to EGFR/Ras/MAPK. Lineage-wide reduction of Hh signaling decreases CySC numbers, while overexpression expands the CySC population. Importantly, CySCs lacking Hh signaling are not maintained, whereas clones with hyperactivated Hh signaling exhibit marked hypercompetitiveness (Amoyel et al., 2013; Amoyel et al., 2014; Michel et al., 2012), supporting a dominant role for Hh in regulating competition.

The Hpo pathway presents a more complex picture. Lineage-wide activation of the pathway, achieved by loss of upstream regulators such as Expanded or Merlin, increases CySC numbers, yet lineage-wide inhibition through dominant-active Merlin produces a similar phenotype (Inaba, Sorenson, Kortus, Salzmann, & Yamashita, 2017; Johnson & Leatherman, 2021; Sun, Zhao, & Wang, 2008). At the clonal level, however, activating the downstream effector Yki confers a competitive advantage, whereas loss of Yki leads to rapid clonal elimination (Amoyel et al., 2014). Taken together, these findings suggest that Yki activity primarily governs competitive behavior rather than core self-renewal.

Collectively, these studies support a model in which JAK/STAT signaling functions as the principal driver of CySC self-renewal, while Hh, Hpo/Yki, and EGFR/Ras/MAPK pathways predominantly modulate competitive interactions within the niche (Fig. 2). Within this framework, activation of the JAK/STAT self-renewal program induces expression of Socs36E, which attenuates MAPK signaling and constrains excessive competitive signaling. By coupling self-renewal inputs to repression of competition-promoting pathways, Socs36E acts as a key integrator of niche-derived signals, ensuring stable coexistence of stem cells within a shared niche.

What happens to the “losing” CySCs?

In classical models of stem cell competition, losing stem cells are thought to be passively displaced from the niche and to subsequently enter a differentiation program. However, recent work in the Drosophila testis suggests that competitive outcomes are more actively regulated. Cell competition studies in developing Drosophila epithelia identified a conserved competitive signaling module involving the Toll (Tl) receptor and the NF-κB transcription factor Relish (Rel), which functions within loser cells to promote their elimination, through apoptosis (Alpar et al., 2018; Meyer et al., 2014). Notably, overt cell death has not been observed in loser CySC clones in the fly testis (Amoyel et al., 2014), raising the question of whether this pathway plays a role in this context. Addressing this, Hof-Michel et al. examined CySC competition in the testis and found that loser clones - defined by reduced Hh or JAK/STAT signaling, or loss of Zfh1 - were underrepresented within the Zfh1-positive stem cell compartment at 5 dpci. Importantly, simultaneous loss of Tl or Rel in these loser clones significantly increased their retention within the Zfh1-positive population at 5 dpci (Hof-Michel, Cigoja, Huhn, & Bokel, 2024). Conversely, ectopic activation of Rel in otherwise competitive ptc-mutant CySCs caused their displacement away from the hub into the second tier of Zfh1-positive cells. Together, these findings identify NF-κB signaling as an effector of loser status in the testis niche.

Future directions

The studies discussed above rely primarily on fixed tissue, an approach that is essential for quantifying clone behavior at the population level and over extended timescales. Nonetheless, growing evidence that multiple signaling pathways influence CySC competitiveness highlights important gaps in our understanding of how these inputs are coordinated. A key future direction will be to determine whether pathways such as Hh, Hpo, EGFR/Ras/MAPK, and NF-κB signaling operate within a genetic hierarchy or function in parallel to converge on common cellular outputs, including niche occupancy, cell proliferation, cell adhesion, and CySC motility. Clarifying how self-renewal cues interface with pathways that promote loser behaviors will be critical for understanding how CySCs integrate permissive and antagonistic signals to determine competitive outcomes within the niche. Additionally, there are other factors that regulate the number of CySCs, including the PDGF/VEGF homolog Pvr (Mues, Hammer, & Leatherman, 2023). It will be important in the future to examine their roles, if any, in CySC-CySC competition.

Fundamental aspects of CySC behavior remain poorly defined, including how CySCs divide, how daughter cells are retained within or excluded from the niche, and whether competitively advantaged CySCs actively displace neighboring somatic cells or instead capitalize on transient access to niche signals. Recent advances in live imaging of the somatic lineage in the Drosophila testis have begun to address these questions (Garcia-Tejera, Tian, Amoyel, Grima, & Schumacher, 2025; Roach, 2025; Roach, Harsh, Sainath, Bach, & Lenhart, 2025). In particular the work from Roach revealed that CySCs are highly motile, are not stably paired with individual GSCs, and frequently remodel their contacts with hub cells and germ cells (Roach, 2025). Live imaging has also captured CySC self-renewal events for the first time, demonstrating that displaced daughters can complete cytokinesis before returning to stem cell positions, with substantial variability in timing. Moreover, CySC daughters exhibit unexpected plasticity, extending protrusions, contacting multiple germ cells over prolonged periods, and often encapsulating single germ cells prior to forming stable cysts - often outside the niche altogether (Roach, 2025) (Kari Lenhart, personal communication). Consistent with this dynamic view, recent expansion microscopy studies challenge the long-held assumption of a fixed 2:1 soma-to-germline ratio within the niche (Ben-Hamo-Arad et al., 2026). Together, these findings underscore the need to integrate cell behaviors revealed by live imaging with molecular models to achieve a mechanistic understanding of somatic stem cell competition in the testis.

CySC-GSC competition in the Drosophila testis

Because space around the hub is limited, CySCs and GSCs actively compete with each other for access to hub-derived self-renewal signals. Indeed, CySCs rendered hypercompetitive through increased Hh, Yki, or MAPK signaling, or through loss of the negative regulator Socs36E, can displace GSCs from the niche, ultimately leading to GSC loss (Fig. 5) (Amoyel, Anderson, et al., 2016; Amoyel et al., 2014; Issigonis et al., 2009). Following displacement, the invading CySCs occupy the positions formerly held by GSCs (Fig. 5). Importantly, reducing the dosage of core cell-cycle regulators such as stg significantly suppresses GSC loss. In fact, CySC clones with an accelerated cell cycle (i.e., overexpressing CycE and Stg) outcompeted wild-type GSCs at the niche in a manner reminiscent of ptc-mutant clones (Amoyel et al., 2014). Taken together, these data indicate that accelerated proliferation of the invading CySC population is a key driver of CySC-GSC competition.

Figure 5. CySC-GSC competition in the adult Drosophila testis.

CySC clones with enhanced competitive fitness (green) preferentially expand within the niche, increasing the likelihood of clonal fixation and the formation of monoclonal CySC pools. Such fitness advantages can arise from elevated Hh, Yki, or MAPK signaling, loss of Socs36E or Abl, or accelerated cell-cycle progression driven by CycE and Stg. As hypercompetitive CySCs and their progeny colonize the CySC pool, they displace wild-type GSCs (pink) from the hub (yellow), leading to progressive GSC loss. Following displacement, invading CySCs occupy positions formerly held by GSCs, reshaping niche composition and stem cell balance.

The role of adhesion in CySC-GSC competition has been reported. One mutant allele of rhea (rhea^6-66), encoding the Drosophila talin that connects integrins to the actin cytoskeleton, partially suppressed outcompetition of GSCs by competitive Socs36E-mutant CySCs (Issigonis, 2009). While this same rhea allele suppressed GSC competition by ptc-mutant CySCs, another rhea allele (rhea¹) did not (Amoyel et al., 2014). In the case of ptc, the requirement for rhea is probably indirect. Rhea/Talin is required to tether the hub to overlying muscle sheath and global loss of Rhea loosens this tethering and permits more stem cells to surround the hub (Tanentzapf et al., 2007). Consistent with this idea, there were more GSCs in testes from rhea¹/+ heterozygous animals compared to controls (Amoyel et al., 2014). Finally, GSC competition by ptc-mutant CySC clones was not perturbed by reducing the genetic dose of α Catenin.

The observation that GSC loss occurs only after hypercompetitive CySCs have largely replaced wild-type CySCs further suggests a hierarchical process in which mutant CySCs first eliminate neighboring somatic competitors before encroaching on the germline compartment (Fig. 5). However, not all mutations that enhance CySC competitiveness trigger CySC-GSC competition: for example, Abl-mutant CySCs exhibit a competitive advantage over wild-type CySCs without affecting GSC number (Stine et al., 2014), indicating that CySC-CySC and CySC-GSC competition can be genetically uncoupled.

Future directions

A major unresolved question is whether hypercompetitive CySCs displace GSCs through an active, directed process or whether they send pro-differentiation signals to GSCs. Discriminating between these models will require direct visualization of competitive interactions in real time. The development of tools that enable simultaneous live imaging of both somatic and germline lineages within the same testis are being developed and will be critical for determining whether colonizing CySCs physically invade, dislodge, or otherwise destabilize resident GSCs, or whether GSC loss precedes CySC occupation of vacated niche space (Roach, 2025) (Kari Lenhart, personal communication).

Beyond imaging, future work should address the molecular mechanisms that distinguish CySC mutations capable of driving CySC-GSC competition from those that selectively affect CySC-CySC interactions. Identifying the specific cellular properties and common target genes that permit CySCs to encroach upon the germline niche will clarify how competitive hierarchies are established across lineages. Finally, it will be important to determine the long-term consequences of CySC-driven GSC displacement for tissue homeostasis, fertility, and aging, and whether similar cross-lineage competitive interactions operate in other stem cell systems. Together, these approaches will help define how competitive behaviors are constrained or permitted across lineage boundaries within shared niches.

GSC-GSC competition in the Drosophila testis

GSCs in the testis undergo oriented mitosis in a stereotyped manner, producing one daughter that remains in contact with the hub and retains stemness and another daughter that is displaced and differentiates (Yamashita et al., 2003). The first observation of replacement of a GSC by neighbors was reported in 2006. Using clonal analysis to mark a wild-type GSC clone and its descendants, these authors found that during aging GSC clones were lost from their niche at a significant rate. However, since the total number of GSCs at the niche in these testes remained high, they concluded that regular GSC loss and replacement by neighbors occurs during aging (Wallenfang, Nayak, & DiNardo, 2006). They speculated that the replacement could occur through symmetric division of GSCs. Extended live imaging of GSCs in the adult Drosophila testis did not observe symmetric GSC divisions but did reveal potential mechanisms through which GSC replacement can occur (Sheng & Matunis, 2011). While 80% of GSC divisions during homeostasis resulted in one stem cell daughter and one daughter cell poised to differentiate, new wild-type GSCs were generated in the niche by “symmetric renewal” in 7% of GSC divisions. Symmetric renewal occurs when the Gb of a GSC-Gb pair swivels back into the niche and becomes a GSC. In addition, this live imaging captured a direct differentiation of a GSC by detachment from the hub in 13% of divisions (Sheng & Matunis, 2011). These data show that stereotyped division orientation (hub-anchored vs. displaced daughter) does not rigidly dictate asymmetric fate. Instead, the niche modulates stem cell behavior depending on tissue needs (steady-state vs. loss/regeneration).

Competition between GSCs that give rise to gametes directly impacts the proportion of alleles transmitted to the next generation. A GSC with a competitive advantage should be selected for and should transmit its alleles at super-Mendelian (> 50%) rates (Otto & Hastings, 1998). Thus, GSC competition has long-term implications for genetic drift and evolution. Despite its importance, GSC competition is an understudied area (reviewed in (Hodge & Bach, 2024)).

chinmo-mutant GSCs outcompete wild-type GSCs for access to the hub.

A putative transcription factor, Chinmo is a nuclear protein expressed in all cell types within the testis stem cell niche, including GSCs (Flaherty et al., 2010). To determine the role of chinmo in GSCs, sparse control or chinmo-mutant GSC clones were induced at equal frequencies, and their behavior was monitored over time at defined times after clone induction. As time progressed, chinmo-mutant GSC clones increasingly dominated the niche, ultimately generating a monoclonal germline (Fig. 6B), a behavior infrequently observed in wild-type GSC clones (Fig. 6A) (Tseng et al., 2022). Quantification of the percentage of GSCs within the niche that were clonal (termed the “clone occupancy”) revealed that chinmo-mutant clones showed a significant increase in occupancy over time, indicating that loss of chinmo confers a competitive advantage. This expansion was not due to over-proliferation, as chinmo-mutant and control GSC clones were present in comparable numbers throughout the experiment. Instead, wild-type neighboring GSCs were preferentially lost in the presence of chinmo-mutant GSCs, resulting in a net decline in GSC number (Fig. 6B).

Figure 6. GSC-GSC competition in the adult Drosophila testis.

(A) Neutral competition. Under steady-state conditions, a labeled GSC clone (gray) may expand or be lost stochastically. Most clones are eliminated over time, whereas rare persisting clones can expand and, in some cases, ultimately dominate the GSC pool, leading to monoclonality.

(B) Biased competition among GSCs. A competitive GSC clone and its progeny (green) remodel the niche microenvironment by secreting Perlecan (Pcan; gray), leading to formation of a “moat” surrounding the hub (yellow) and recruitment of Laminin (Lan; black) from the adjacent testis muscle sheath. This altered extracellular matrix selectively destabilizes wild-type GSCs (pink), promoting their loss from the niche. In contrast, competitive GSCs are retained through upregulation of Dystroglycan (Dg; orange) and βPS integrin (βPS; red) (inset). Over time, this biased competition results in monoclonal replacement of the germline and biased transmission to offspring. Hypercompetitive GSCs can arise through clonal loss of Chinmo or of Imp, or through clonal overexpression of H.

chinmo-mutant GSCs secrete Perlecan to remodel the niche, causing wild-type GSC loss.

chinmo-mutant GSCs gained a competitive advantage by secreting the extracellular matrix component Perlecan (Pcan) into the testis lumen, thereby remodeling the niche microenvironment. Secreted Pcan accumulated around the hub and promoted recruitment of Laminin (Lan) from the basal lamina of the surrounding muscle sheath, resulting in formation of a characteristic “moat” encircling the hub (Tseng et al., 2022). These moats were not present in testes containing control clones (Fig. 6B). Importantly, Pcan secretion was required for this competitive behavior: knockdown of Pcan in chinmo-mutant clones abolished their enhanced niche occupancy and prevented displacement of neighboring wild-type GSCs.

chinmo-mutant GSCs remain in the altered niche by upregulating ECM-binding proteins

chinmo-mutant GSCs persisted in the remodeled niche by upregulating ECM-binding proteins (Fig. 6B) (Tseng et al., 2022). Mutant clones exhibited elevated expression of Dystroglycan (Dg) and βPS integrin at the GSC-niche interface, enabling them to maintain niche occupancy even as wild-type GSCs were lost. Both Dg and βPS were required for this advantage, likely through enhanced adhesion to the niche ECM. Depletion of Dg in chinmo-mutant clones reduced their clone occupancy. Moreover, supplying wild-type GSCs with additional Dg protected them from competitive elimination, demonstrating that the adhesion advantages acquired by chinmo-mutant GSCs are both necessary and sufficient to confer “winner” status (Tseng et al., 2022).

GSC competition causes biased inheritance

Because most testes containing chinmo-mutant GSC clones ultimately become homozygous mutant, the authors hypothesized that the chinmo⁻ allele would be inherited at super-Mendelian frequencies (Fig. 6B) (Tseng et al., 2022). To test this, they developed an inheritance assay in which GSC clones were generated in males carrying either a chinmo⁺ or chinmo⁻ allele in trans to a ubiquitously expressed GFP-marked chromosome that could be followed in the next generation. After allowing competition to proceed, these males were crossed to virgin females. Offspring from males with chinmo⁺ GSC clones displayed the expected Mendelian ratio: 50% inherited the chinmo⁺ allele and 50% inherited the GFP-marked chromosome. In contrast, 65% of the offspring from males carrying chinmo¹ (a strong hypomorph) GSC clones inherited the chinmo¹ allele, with only 35% inheriting the GFP-positive chromosome, a significant deviation from Mendelian inheritance.

chinmo-mutant GSC coopt a natural aging program for their competitive behavior

The authors investigated whether age-related changes in the Drosophila testis niche resemble the competitive phenotypes seen in chinmo-deficient GSCs (Fig. 7). They found that Chinmo protein levels declined progressively in GSCs during adulthood, coinciding with a significant reduction in total GSC number and the appearance of a characteristic “moat” structure composed of Pcan and Lan in aged testes. These age-associated changes, including decreased Chinmo, reduced GSC number, the presence of the moat, and increased Dg at the GSC-niche interface, mirrored phenotypes caused by induction of chinmo-mutant GSC clones in young males.

Figure 7. Model for age-dependent regulation of the BMP-Hairless-Imp-Chinmo axis.

(Left) In a young testis, hub cells (yellow) produce high levels of BMP ligands (gray circles), activating BMP receptors on an adjacent GSC (dark pink). Robust BMP signaling represses the transcriptional co-repressor Hairless (H), permitting high Imp expression and promoting accumulation of Chinmo. Elevated Chinmo represses Pcan and Dg, maintaining a youthful GSC state and preserving niche architecture.

(Right) In an aged testis, BMP production by hub cells (light yellow) and BMP signaling in a GSC (light pink) declines. Reduced BMP signaling relieves repression of H, leading to increased H levels, repression of Imp, and a consequent decrease in Chinmo. Loss of Chinmo derepresses Pcan and Dg, resulting in Pcan secretion by GSCs and increased Dg accumulation at the GSC-niche interface. Secreted Pcan accumulates around the hub and recruits Lan from the muscle sheath basal lamina, thereby remodeling the niche microenvironment during aging.

To test whether maintaining Chinmo could prevent these changes, the authors overexpressed Chinmo in GSCs throughout adulthood. Sustained germline Chinmo expression effectively suppressed all major aging phenotypes, including the moat, elevated Dg, and GSC loss, whereas control overexpression of a neutral protein LacZ had no effect. These findings demonstrate that declining Chinmo levels intrinsically drive aging of the testis stem cell niche and that chinmo-mutant GSCs in young males exploit this natural aging program to remodel the niche to their competitive advantage (Tseng et al., 2022).

GSC clones that ultimately reduce Chinmo levels cause biased competition

Recent work has defined a niche-driven regulatory axis linking GSC aging and competition through BMP-H-Imp-Chinmo signaling (Fig. 7) (Zheng et al., 2025). In young testes, hub-derived BMPs repress the transcriptional co-repressor H in GSCs, permitting Imp expression and sustaining Chinmo to preserve GSC number. Age-associated decline in BMP production by hub cells led to reduced BMP signaling in GSCs. This in turn allowed H to accumulate, causing repression of Imp, and reduction of Chinmo, driving stem cell and niche deterioration. Disruption of this axis also biased stem cell competition, as GSC clones with reduced Imp or elevated H progressively outcompeted wild-type neighbors and dominated the niche in manner that is indistinguishable from chinmo-mutant GSC clones. Thus, a reduced BMP-H-Imp-Chinmo axis both accelerates GSC aging and confers a competitive advantage on stem cells with low Chinmo levels (Zheng et al., 2025).

What other genes regulate competition between GSCs in the testis?

GSC clones carrying mutations in the lipase Brummer expand to occupy a greater fraction of the stem-cell pool than control clones (Chao et al., 2022). It remains to be determined whether brummer-mutant GSCs gain their advantage through mechanisms like those used by chinmo-mutant GSCs.

Future directions

Together, these studies establish GSC-GSC competition as a dynamic and biologically consequential process shaped by intrinsic stem cell state, niche architecture, and age-associated signaling decline. A key priority moving forward will be to identify additional molecular determinants of competitive fitness in GSCs. In this context, forward genetic screens in the testis designed to isolate mutations that enhance GSC fitness represent a powerful and unbiased strategy for discovering new regulators of GSC competition. In parallel, developing quantitative models of GSC competition, analogous to those used to describe neutral drift and biased competition among CySCs, will be essential for distinguishing stochastic from deterministic behaviors and for predicting long-term clonal outcomes. Such approaches should also clarify whether additional modes of competition operate in the testis beyond niche remodeling, including mechanisms involving metabolism, stress resistance, or cell–cell signaling. Finally, defining how competitive interactions are initiated and resolved in real time, particularly during aging, regeneration, or environmental stress, will benefit greatly from emerging live-imaging and single-cell approaches.

Competition between dedifferentiating germ cells and resident stem cells

Dedifferentiation refers to the process by which lineage-committed cells revert to a stem or progenitor state within the same developmental pathway. In the Drosophila testis, experimental perturbations that transiently disrupted reception of self-renewal cues or temporarily overexpressed a differentiation factor forced all GSCs to differentiate. When this occurred, CySCs moved into the positions vacated by the GSCs (Brawley & Matunis, 2004; Sheng et al., 2009). Once the perturbations were removed, interconnected spermatogonia fragmented into individual cells that actively migrated back toward the hub. These reverting germ cells displaced resident CySCs, re-established GSC identity by activating JAK/STAT signaling, and suppressed differentiation programs (Brawley & Matunis, 2004; Sheng et al., 2009). Live imaging revealed that dedifferentiating spermatogonia extended dynamic, actin-rich protrusions as they approached the niche (Sheng & Matunis, 2011), supporting a model in which reverting germ cells gain a competitive advantage through enhanced adhesive or invasive behavior that allows them to reclaim hub contact from CySCs.

Germline dedifferentiation is not limited to experimentally manipulated testes but also occurs during aging and in response to physiological stresses like starvation and mating (Fig. 8). In this case, the dedifferentiating germ cells must outcompete resident CySCs and resident GSCs at the niche (Fig. 8). Live-imaging results indicated that in young testes, the rate of dedifferentiation was very low but significantly increased after mating (Roach & Lenhart, 2024; Sheng & Matunis, 2011). Lineage-tracing experiments demonstrated that spermatogonial reversion contributes increasingly to the GSC pool with age, such that approximately 40% of GSCs in older testes originated from dedifferentiated germ cells (Bener, DiPippo, Slepchenko, & Inaba, 2025; Cheng et al., 2008; Herrera & Bach, 2018). This age-associated dedifferentiation was stimulated by Jun N-terminal kinase (JNK) signaling, although whether JNK primarily promoted cyst breakdown, niche re-entry, or competitive displacement of resident stem cells remains unresolved (Herrera & Bach, 2018). Notably, GSCs generated through dedifferentiation exhibited higher proliferative capacity and produced more daughter cells than long-lived resident GSCs within the same testis. Thus, dedifferentiation acts as a rejuvenation mechanism, replenishing the niche with competitively superior stem cells and partially counteracting age-related declines in testis function (Herrera & Bach, 2018).

Figure 8. Dedifferentiating germ cells outcompete wild-type GSCs and wild-type CySCs.

A 4-cell spermatogonial cyst fragments, releasing differentiating germ cells (light purple) from their syncytial neighbors. One spermatogonial cell migrates back toward the hub. To gain niche access, the dedifferentiating germ cell must outcompete resident stem cells, either a wild-type GSC (purple), a wild-type CySC (blue), or both. Upon successful niche entry, the cell is reclassified as a GSC derived from dedifferentiation (red).

Future directions

Several key questions remain regarding how dedifferentiation is regulated and integrated into niche homeostasis. First, the molecular determinants that confer competitive superiority on dedifferentiated GSCs are poorly defined; dissecting how adhesion, cytoskeletal dynamics, metabolism, and signal responsiveness are rewired during reversion will be essential. In particular, it will be important to determine whether actin-rich protrusions and migratory behaviors are causally required for niche re-entry or instead reflect a broader competitive state. Second, the role of JNK signaling needs to be resolved with temporal and cell-type specificity to distinguish its functions in cyst fragmentation, directed migration, niche invasion, and displacement of resident stem cells. Third, how dedifferentiation interfaces with age-related changes in the niche, such as declining self-renewal signals or altered behavior of resident stem cell populations, remains largely unexplored. Finally, whether repeated cycles of dedifferentiation impose long-term costs, such as increased mutational burden or altered lineage fidelity, is unknown. Addressing these questions will clarify whether dedifferentiation represents a universally beneficial rejuvenation strategy or a trade-off that preserves tissue function at the expense of long-term genomic or cellular stability.

Stem cell competition in mammalian systems

Health outcomes of somatic stem cell competition

Advances in DNA sequencing over the past decade have revealed that adult stem cells in many human tissues acquire somatic mutations continuously after birth, accumulating damage at an approximately linear pace across lifespan, with pronounced tissue-specific differences (Manders, van Boxtel, & Middelkamp, 2021; Moore et al., 2021). Notably, this mutational accumulation is largely uncoupled from stem cell division and is also detected in long-lived, postmitotic cells such as neurons (Abascal et al., 2021). Because stem cells persist for decades, the aggregate mutational load across the body becomes extensive, such that over a lifetime, much of the genome is altered somewhere within the overall stem cell population. As a result, aging humans become genetic mosaics in which many ostensibly normal, non-malignant tissues are populated by positively selected clones carrying canonical cancer driver mutations (reviewed in (Derks & van Boxtel, 2023; Kakiuchi & Ogawa, 2021)). In epithelial compartments, these mutant stem cells arise within a field of normal neighbors and are therefore subject to immediate local competition for limited space, with outcomes that can either restrain tumor formation or, conversely, promote oncogenic expansion.

In stratified epithelia, the architecture of the basal layer permits extensive lateral spread of mutant clones, creating a competitive landscape in which clonal expansion can either promote or restrain tumor initiation (reviewed in (Fowler & Jones, 2022)). In the normal esophageal epithelium, clones mutant for the tumor suppressor NOTCH1 dominate in terms of tissue coverage, such that when combined with loss of heterozygosity (LOH), most of the cells in the esophagus have lost both NOTCH1 alleles by midlife (Abby et al., 2023; Fowler et al., 2021). Strikingly, despite their strong competitiveness in normal tissue, NOTCH1-mutant clones are rarely represented in esophageal cancers, accounting for fewer than 10% of tumors (Cancer Genome Atlas Research et al., 2017). This underrepresentation suggests that NOTCH1 loss may actively impede malignant progression. By contrast, alterations in the tumor suppressor TP53 show a different pattern: TP53 mutations accompanied by LOH are nearly universal in esophageal squamous carcinomas; however, in normal epithelium, TP53-mutant clones are far less prevalent. Although heterozygous TP53 mutations expand to occupy ~10% of the tissue by midlife and up to ~30% in elderly individuals, clones with complete TP53 loss remain rare in non-malignant tissue (Hall, Jones, & Hall, 2019; Martincorena et al., 2018; Yokoyama et al., 2019), highlighting a sharp distinction between competitive success in normal epithelia and oncogenic potential.

In the mouse intestinal epithelium, ISC clones that inactivate the tumor suppressor Adenomatous polyposis coli (Apc) or that overexpressed dominant-active forms of the proto-oncogenes Kras (Kras^G12D) or PIK3CA (PIK3CA^H1047R) have an elevated probability to achieve crypt fixation, shifting stem-cell dynamics away from neutral toward biased competition (Snippert et al., 2014; Vermeulen et al., 2013; Yum et al., 2021). In fact, these mutant ISCs behave as supercompetitors, and they dominate the local milieu by promoting differentiation of neighboring wild-type stem cells and their exit from the niche, thereby freeing space for mutant clonal expansion (Flanagan et al., 2021; van Neerven et al., 2021; Yum et al., 2021). Apc-deficient ISCs secrete WNT antagonists that drive neighboring wild-type ISCs out of the stem cell compartment (Flanagan et al., 2021; van Neerven et al., 2021). Disrupting this non-cell-autonomous mechanism, either by deleting the WNT inhibitor Notum or by enhancing WNT signaling in wild-type ISCs through pharmacological inhibition of GSK3β, attenuated the competitive advantage of Apc-mutant cells, reduced crypt fixation, and suppressed adenoma formation. Kras^G12D and PIK3CA^H1047R mutant ISCs influenced stem-cell competition by promoting secretion of factors that activate BMP signaling. Because BMP signaling antagonizes WNT-dependent self-renewal and proliferation in the intestinal epithelium, this shift reduces the pool of wild-type competitors (Yum et al., 2021). This niche remodeling increases the likelihood that the supercompetitor stem cell will achieve complete crypt fixation and initiate adenoma formation.

In the aging human bone marrow, hematopoietic stem cells (HSCs) can undergo clonal hematopoiesis, a state in which a single mutant clone disproportionately contributes to the stem-cell pool. This phenomenon occurs in ~10% of individuals over age 70 and is a major risk factor for hematologic malignancies and inflammatory disease (Genovese et al., 2014; Jaiswal et al., 2014; Watson et al., 2020). Large-scale surveys reveal that most clonal events arise from mutations in the DNA methyltransferase DNMT3A or in the methylcytosine dioxygenase TET2 (Jaiswal & Ebert, 2019; Jakobsen et al., 2024). When such mutations are present in ≥4% of nucleated blood cells without overt disease, the condition is termed clonal hematopoiesis of indeterminate potential (CHIP). Studies of CHIP have illuminated mechanisms of stem-cell competition: in mouse transplantation assays, DNMT3A - or TET2-mutant HSCs consistently outcompeted wild-type HSCs (Challen et al., 2011; Ko et al., 2011), while in humans these mutants showed enhanced resistance to age-associated inflammatory stress (Caiado et al., 2023; Hormaechea-Agulla et al., 2021; Jakobsen et al., 2024; Zioni et al., 2023). For DNMT3A mutants, increased chromatin accessibility and activation of growth-promoting programs further reinforce competitive fitness (Nam et al., 2022). Additional mutations confer a competitive advantage to HSCs, including loss of TP53 and gain of the Janus kinase JAK2 (Bondar & Medzhitov, 2010; Kralovics et al., 2005). Large-scale sequencing of healthy individuals has revealed that mutations in HSCs are under positive natural selection (Watson et al., 2020). These results support a model in which ongoing acquisition of mutations and selection for advantageous variants underlies the increasing detectability of clonal hematopoiesis with age, linking somatic evolution in normal tissues to cancer risk.

Collectively, these findings indicate that cancer risk arises not merely from the progressive accumulation of mutations, but from selective forces acting on mutant stem cells within competitive niches. Age-related increases in cancer susceptibility reflect prolonged periods of competitive selection, during which mutant stem cells with enhanced fitness gradually replace their normal counterparts. This perspective recasts carcinogenesis as an evolutionary process unfolding within stem-cell compartments and underscores stem-cell competition as an early mechanistic link between aging, clonal expansion, and cancer initiation (see (Henry & DeGregori, 2025)).

Neutral competition between mouse spermatogonial stem cells

Mouse spermatogenesis provided one of the earliest demonstrations of population-asymmetric self-renewal, alongside the studies in epithelia (Klein et al., 2010; Nakagawa, Nabeshima, & Yoshida, 2007; Nakagawa, Sharma, Nabeshima, Braun, & Yoshida, 2010). Clonal analyses of sparsely marked mouse spermatogonial stem cells (SSCs) demonstrated that individual stem cells follow markedly divergent trajectories over time (Hara et al., 2014). Some labeled SSCs underwent symmetric self-renewing divisions that increased their representation within the stem-cell pool, whereas others generated only differentiating progeny and were ultimately lost. Despite this pronounced heterogeneity at the single-clone level, the collective behavior of the population was tightly regulated, with self-renewal and differentiation occurring at equal rates, thereby maintaining tissue homeostasis (Hara et al., 2014; Klein et al., 2010). Notably, the distribution of clone sizes and fates closely matched predictions from a stochastic model in which all SSCs are functionally equivalent but make probabilistic fate decisions. These findings show that SSCs are maintained by population asymmetry through neutral competition.

While the spatial organizing of the closed-niche system of the Drosophila testis dictates that stem cells compete for niche access, there must be a different paradigm at play in SSC competition an open-niche system like the mammalian testis (reviewed in (Kitadate & Yoshida, 2022)). Along the seminiferous tubules, SSCs are dispersed and migratory, but they maintain quite constant average densities even after partial ablation (Kanatsu-Shinohara et al., 2003; Kitadate et al., 2019). This robust recovery to a defined density indicates that population size is regulated by self-organizing mechanisms operating at the population level. A key mechanism underlying this regulation is termed “mitogen competition”. Both GDNF - signaling through the GFRα1/RET receptor complex - and Fibroblast growth factor (Fgf) - signaling through the Fgf receptor - can act as self-renewal cues for mouse SSCs (Kanatsu-Shinohara & Shinohara, 2013). As SSCs migrate along the basement membrane, they compete for a limited supply of FGF produced by somatic support cells (Kitadate et al., 2019). SSCs that acquire higher levels of FGF are biased toward self-renewal, whereas those receiving less signal are more likely to differentiate. Competition for this shared mitogen pool is sufficient to stabilize stem-cell density and explains the local enrichment of SSCs near FGF sources as an emergent outcome, without invoking fixed niches or chemoattraction. It will be important in the future to determine whether this mechanism is operative in other open-niche systems.

Human SSC Competition and Paternal Age Effect Disorders

Advances in sequencing have led to new insights into mutations in human germ cells (reviewed in (Wood & Goriely, 2022)). Each newborn carries on average ~60 de novo point mutations (DNMs) (ranging from ~30–90 depending on parental age), corresponding to a human germline mutation rate of ~1.2 × 10⁻⁸ per nucleotide per generation. The total number of DNMs increases steadily with parental age and is strongly biased toward the paternal germline: ~80% of DNMs arise on the paternally inherited allele, and paternal age at conception is the dominant determinant of DNM burden (Goldmann, Veltman, & Gilissen, 2019; Goldmann et al., 2016; Jonsson et al., 2017; Kaplanis et al., 2022; Kong et al., 2012; Rahbari et al., 2016). On average, ~1–2 additional DNMs are added to the offspring genome for each year of paternal aging (Acuna-Hidalgo, Veltman, & Hoischen, 2016; Jonsson et al., 2017; Kong et al., 2012; Rahbari et al., 2016). Notably, this age-related accumulation occurs even though germ cells harbor mutation rates several orders of magnitude lower than those of somatic tissues, implying the existence of specialized mechanisms that preserve germline genome integrity, potentially including pervasive testis transcription that facilitates DNA damage detection and repair. The paternal age effect (PAE) on disease risk is therefore causally linked to the gradual accumulation of DNMs in SSCs over time. Classic PAE disorders arise from specific point mutations in genes such as FGFR2, FGFR3, RET, PTPN11, and HRAS, causing conditions including Apert, achondroplasia, multiple endocrine neoplasia, Noonan syndrome, and Costello syndrome that are termed rasopathies (Fig. 9) (Wood & Goriely, 2022). Many of these mutations are thought to act through “selfish selection,” whereby rare gain-of-function mutations arise spontaneously in SSCs, confer a competitive advantage, and clonally expand within seminiferous tubules as men age, leading to marked enrichment of mutant sperm and increased transmission risk (Choi, Yoon, Calabrese, & Arnheim, 2008, 2012; Giannoulatou et al., 2013; Goriely et al., 2009; Goriely, McVean, Rojmyr, Ingemarsson, & Wilkie, 2003; Goriely et al., 2005; Maher et al., 2016; Shinde et al., 2013; Yoon et al., 2013). Although these strongly selfish mutations become more frequent in the sperm of older men, they typically cause severe disease and reduced reproductive fitness in offspring, limiting their persistence across generations.

Figure 9. Model for human paternal age effects (PAEs).

(A) A wild-type spermatogonial stem cell (SSC; pink) divides to generate daughter cells that differentiate into mature sperm. A de novo mutation in a PAE-associated gene can arise in an SSC (green). This mutant SSC undergoes clonal expansion, producing a disproportionately large number of sperm relative to wild-type SSCs. As a result, sperm derived from a single mutant SSC are more likely to fertilize an egg than sperm produced by an individual wild-type SSC.

(B) Schematic of a germ cell highlighting known selfishly selected genes (red), which cluster within the receptor tyrosine kinase (RTK)/Ras/MAPK signaling pathway. In this context, the relevant RTKs include FGFR and RET.

A and B were adapted from Wood and Goriely, Fertil Steril 2022 Vol. 118 Issue 6 Pages 1001–1012 10.1016/j.fertnstert.2022.10.017

There are some mouse models of human PAE. For example, Hras^G12V induction after fertilization recapitulates human Costello syndrome in mouse offspring (X. Chen et al., 2009). However, conditional induction of Hras^G12V in undifferentiated murine spermatogonia that includes SSCs did not provide a competitive advantage to SSCs over a period of one year (Yamada, Cai, Martin, N’Tumba-Byn, & Seandel, 2019). Thus, there is not yet a mouse model for studying PAE in vivo. However, SSC cell lines have provided experimental results supporting the central hypothesis of positive selection propagation of SSCs in mammals. Adult mouse SSC lines ectopically expressing Apert syndrome-associated human FGFR2 mutation (S252W) exhibited significantly greater stem cell activity than those with ectopic wild-type human FGFR2 in transplantation assays (Martin, Assif, Gilbert, Wijewarnasuriya, & Seandel, 2014). S252W-expressing SSCs did not show any proliferative advantage when cultured alone in vitro in standard media but exhibited increased competitiveness in the presence of wild-type SSCs. S252W-expressing SSCs showed enhanced sensitivity to FGF, as evidenced by enhanced MAPK signaling with low dose FGF, suggesting an increased self-renewal capability of S252W SSCs, when growth factors are scarce.

Possible link between human germ cell tumors and SSC competition

Spermatocytic tumors (SpTs), previously referred to as spermatocytic seminomas, comprise a rare subset of testicular germ cell tumors (Carriere, Baade, & Fritschi, 2007; Goriely et al., 2009). Representing roughly 1% of testicular cancers and most often arising in older men (median age of diagnosis is 54 years), SpTs grow slowly and are clinically indolent (Carriere et al., 2007; Lim et al., 2011; Reuter, 2005). The cell of origin for SpTs is adult germ cells, probably spermatogonia, and it has been hypothesized that SpTs arise through “selfish” clonal expansion driven by competitive growth advantages within the germline stem cell compartment (Maher, Goriely, & Wilkie, 2014; Rajpert-De Meyts et al., 2003). While the majority of SpTs follow a benign course, a small but clinically significant fraction (5–6%) progress to sarcomatous transformation, a transition associated with metastatic spread, treatment resistance, and poor prognosis (Wetherell, Lawrentschuk, & Gyomber, 2013). Elucidating the cellular origins and competitive mechanisms that drive SpT development will therefore be essential for improving diagnosis and therapeutic strategies for aggressive disease variants.

Conclusions

Work in the Drosophila testis has established stem cell competition as a property of a shared niche wherein access to hub contact and short-range signals is continuously renegotiated. In the somatic compartment, CySCs provide a clear example of neutral competition in an adult stem cell system: a pool of equipotent CySCs maintained by stochastic loss and neighbor replacement. Multiple genetic perturbations bias stem cell competition by shifting the probability of replacement toward the mutant lineage. In most cases, this advantage is driven by increased proliferation, which is both necessary and sufficient to confer niche dominance. Although adhesion can modulate specific contexts, increased adhesive capacity alone is generally not sufficient to drive CySC dominance, reinforcing proliferation as the most consistently supported driver of CySC-CySC competition. Importantly, competitive outcomes are not purely passive as new evidence suggests that elimination from the niche might be instructed, even in the absence of overt apoptosis, adding a mechanistic layer to the classic model of loser displacement from the niche followed by differentiation.

Mechanistically, the best-defined examples of biased GSC-GSC competition emphasize a distinct strategy from the somatic lineage. Winners can remodel the niche microenvironment via ECM secretion and then maintain occupancy by upregulating ECM-binding proteins. Thus, while CySC competition is often explainable as probabilistic bias towards replacement, GSC competition currently stands as a niche-restructuring problem in which winners can co-opt a natural aging program to destabilize neighbors, with super-Mendelian inheritance as a direct readout of competitive success.

Major opportunities for the field are to bring the modeling rigor of CySC studies to the germline in order to define neutral baselines, replacement rules, and quantitative signatures of bias and to test whether changes in stem cell behaviors affects to organism-level outputs like fertility and allele transmission to the next generation. Many of the most important unresolved questions in stem cell competition in the fly testis are inherently dynamic and will be transformed by continued advances in long-term live imaging and simultaneous visualization of somatic and germline behaviors in the same testis.