A complex systems approach to mosaic loss of the Y chromosome

Abstract

Mosaic loss of Y chromosome (mLOY) is an acquired condition wherein a sizeable proportion of an organ’s cells have lost their Y. Large-scale cohort studies have shown that mLOY is age-dependent and a strong risk factor for all-cause mortality and adverse outcomes of age-related diseases. Emerging multi-omics approaches that combine gene expression, epigenetic and mutational profiling of human LOY cell populations at single-cell levels, and contemporary work in in vitro cell and preclinical mouse models have provided important clues into how mLOY mechanistically contributes to disease onset and progression. Despite these advances, what has been missing is a system-level insight into mLOY. By integrating the most recent advances in wide-ranging aspects of mLOY research, we summarize a unified model to understanding the cause and consequence of mLOY at the molecular, cellular, and organismal levels. This model, referred to as the “Unstable Y Cascade model,” states that (i) the rise and expansion of LOY result from interaction by the inherently unstable Y, germline genetic and epigenetic variants, and numerous cell-intrinsic and external factors; (ii) LOY initiates genomic, epigenomic, and transcriptomic alterations in X and autosomes, thereafter induces a cascade of tissue-specific cellular alterations that contribute locally to the onset and progression of diseases; and (iii) LOY cells exert paracrine effects to non-LOY cells, thereby amplifying LOY-associated pathological signaling cascades to remote non-LOY cells. This new model has implications in the development of therapeutic interventions that could prevent or delay age-related diseases via mitigating mLOY burden.

Introduction

Hematopoiesis contributes to daily production of billions blood cells from hematopoietic stem cells (HSCs). The unavoidable intrinsic and extrinsic stresses introduce de novo mutations ranging in size from single-nucleotide variants (SNVs) to chromosome alterations (CAs) in HSCs [1]. If a mutation confers high fitness, the mutated HSCs out-compete wildtype HSCs and produce a pool of progeny harboring the same mutations. This process, created genomic mosaicism in human blood, is termed as clonal hematopoiesis (CH) [2]. Large cohort studies have demonstrated that CH, either derived by pathogenic SNVs or CAs, is age-dependent in healthy individuals [2]. SNVs present at ≥ 2% variant allele fraction in healthy individual with otherwise normal hematologic characteristics is referred as CH of indeterminate potential (CHIP). CAs can be divided into three major subtypes, namely, autosomal CAs, loss of Y chromosome (LOY), and loss of X chromosome (LOX) [1] (Fig. 1).

Fig. 1.

Clonal hematopoiesis in men. Common genetic variants and environmental factors facilitate the acquisition of cells bearing chromosomal alterations and single nucleotide variants in hematopoietic stem cells (HSCs) throughout lifespan. Depending on the degree of selective advantage conferred by each mutation, clones with fitness-enhancing mutations (e.g., loss of Y, TET2 and DNMT3A mutations) will give rise to large mutational clones in advanced age through a process of clonal hematopoiesis and fitness-neutral clones (e.g., loss of X) will be vanished. CH gives rise to a pool of mutant peripheral hematological cells, including lymphocytes, macrophages, and neutrophils. These mutant blood cells are circulated into other tissues and organs via the circulatory system

The initial belief that mLOY is neutral to man health has shifted with the recent epidemiological, multi-omic, and modeling-based studies, which have provided unprecedented insights into a causal connection between mLOY and a broad set of chronic diseases in men [3] and mouse models [4, 5]. One of the biggest challenges in this field is to systematically evaluate the precise causes and consequences of mLOY. Gaining system-level insights to these issues could shed light on fundamental mechanisms in sex differences of human aging and disease [6]. Thus, a systematic approach that links mLOY to biological and subclinical changes and eventually to disease manifestation is required. Building on data from epidemiological and multi-omic studies in humans and mechanistic studies in mouse models, we formulate a novel unstable Y cascade (UYC) model, summarize implications of the UYC model, and discuss open challenges.

Overview of the UYC model

The UYC model (Fig. 2) starts from the rise and expansion of LOY, which are thought to result from interactions between the unstable Y and many other factors. Upon on LOY, expression of all Y-linked genes is lost. Due to loss of single major genes or cumulative effect of loss of multiple minor genes, LOY may induce multi-omic alterations, such as genomic instability (GIN), DNA hypomethylation, and altered expression of genes in X and autosomes. These multi-omic alterations would lead to alterations in molecular signaling and cellular homeostasis, such as increased fitness, pro-inflammatory signaling, impaired immunosurveillance, and myeloid-biased HSC differentiation. Over time, LOY cells expand gradually, driven either by cell-intrinsic mechanisms or by extrinsic stresses. In tissues with mLOY, LOY cells may impact the molecular signaling and cellular homeostasis of non-LOY cells via paracrine signaling cascade. Depending on the cell lineages and cell fractions affected by LOY, these molecular and cellular effects are further relayed to different pathophysiological phenotypes, highlighted by increased transformation rate, cardiac fibrosis rate, immune escape potential, and neuroinflammation level. Finally, these cellular phenotypes lead to increased risk of all-cause mortality and a myriad of age-related diseases in mLOY carriers.

Fig. 2.

The unstable Y cascade (UYC) model. This cascade begins with the unstable Y and can be divided into four stages. In the first stage, the rise and expansion of loss of Y (LOY) result from interactions among the unstable Y, genome-wide genetic and epigenetic variants, and numerous cell-intrinsic and external factors. In the second stage, LOY induces multi-omics alterations in X and autosomes, including genomic, epigenomic, and transcriptomic alteration. In the third stage, multi-omics alterations induce a myriad of tissue-specific cellular alterations. Although some of cellular phenotype alterations are based on correlations, the ongoing development of mouse models is expected to deepen our understanding of how mLOY is involved in different disease phenotypes at the cellular level. Meanwhile, LOY cells exert paracrine effects to non-LOY cells, thereby amplifying LOY-dependent cellular signaling cascades to non-LOY cells and/or to remote non-LOY tissues. In the fourth stage, altered cellular functions contribute in concert to all-cause mortality, aging, and the onset and progression of diseases. Importantly, the full clinical potential of mLOY has yet to be unleashed

The UYC model not only maximizes what can be learned from existing data in mLOY, but also creates a unified framework for studying the biological cause and clinical significance of mLOY. The UYC model applies to mLOY occurring in a range of contexts, including hematopoietic system, solid tumors, and healthy solid tissues. In the following sections, we will focus on the synthesis of the latest discoveries that could provide supports to the UYC model.

The rise and clonal expansion of mLOY

Why some individuals, but not others, exhibit mLOY in the blood in the late life stages? The answer is complicated and primarily depends on the interactions among Y chromosome, genetic and epigenetic risk loci, aging, and environmental stressors.

Human Y is inheritably unstable and prone to be lost

Mammalian Y is genetically highly degenerated: after being diverged with X chromosome (X) from an autosome pair, Y stopped to recombine with X and lost over 90% of its genes. Y sequences are divided into small pseudoautosomal regions (PARs, comprising 5% of Y’s length) and male-specific region of Y (MSY) [7]. Genes located in Y-PARs have homologies in the X-PARs, forming X–Y gene pairs [8]. MSY does not pair with X during meiosis and contains massive palindromic sequences, repetitive sequences, and heterochromatins [8]. Only recently, advances in sequencing and assembling technologies have enabled accurate sequencing of human Y [9]. Although there are extensive variations in size and structure of human Y chromosomes [10], copy number variation on palindromic Y regions has little impact on 341 traits among Icelandic men [11].

A recent study [12] using data of clone size in the blood from approximately 500,000 UK Biobank (UKB) participants to estimate mutation rates and fitness consequences of mosaic CAs was done. The authors found that most CAs have moderate to high fitness effects but occur at a low rate. Two notable exceptions are LOY and LOX, which were estimated to be 1000-fold higher in mutation rates than autosomal CAs [12]. These results should be validated experimentally. In supporting of this, at least 62 independent LOY events were acquired in the blood cells of a 82-year-old male [13]. Why LOY is continuously acquired throughout life? At least two inherent properties make human Y prone to be lost.

The first property is that human Y centromere lacks CENP-B box. CENP-B is a highly conserved centromere protein contributing to centromere strength and the rate of faithful chromosome segregation [14]. Despite centromeres are essential for faithful chromosome segregation, centromeres are rapidly evolving and exhibit chromosome-specific variations in structure [15]. The binding sites for CENP-B, the so-called CENP-B boxes, present in centromeres of a wide range of mammalian species but not detected at the centromeres of virtually all mammalian Y [16]. Absence of CENP-B decreases the strength of Y centromere, therefore making the Y prone to missegregation [17].

The second property is that human Y tends to form structural aberrations [18]. The isodicentric Y is a mirror-imaged Y with two centromeres. When DNA double-strand breaks (DSBs) occur in MSY, they are occasionally repaired by crossover homologous recombination between opposing arms of MSY palindromes from two sister chromatids, which give rise to isodicentric Y [19]. Due to the presence of two functional centromeres, isodicentric Y is mitotically unstable, particularly those with a long intercentromeric distance, leading to LOY [19]. In addition, large and submicroscopic Yq deletions, ring Y, and Y isochromosome are also associated with mLOY in blood cells [20–22].

In addition, it was long hypothesized that, telomeres of human Y are either shorter or shortened at a higher rate than autosomes and X, thereby becoming unstable faster and contributing to frequent LOY with aging [23]. However, evidence for this hypothesis is limited. Although Yp in HG002 reference genome has a significantly shortened telomere length distribution than the grand mean [24] and the attrition rate of Yp telomere is faster than the average telomere attrition rate [25], Y does not have the overall shortest telomere or highest telomere attrition rate among human chromosomes [24].

For mammalian autosomes, there is a chromosome-size-dependent risk of segregation errors during meiosis [26] and mitosis [27]. Using cytogenetic and array-based comparative genomic hybridization data in 43,205 human tumors, it has found that small chromosomes (including Y) are lost more readily than large ones [28]. Thus, it is possible that small size is a potential determinant for LOY. Since Y is gene poor and dispensable, one may argue that LOY is a continuous but clinically irrelevant event during tumor evolution over time. A recent study revealed that, in lung adenocarcinoma, LOY does not co-occur with loss of chromosome 21, which is comparably small and gene-poor to Y [29]. Thus, whether small size and gene poorness are primary determinants for LOY remains a puzzle.

Together, all these structural features make Y be the most unstable chromosome in human, both in short terms (mitotically unstable) or long terms (evolutionarily degenerative). The unstable nature of Y provides a logical explanation for frequent LOY in human somatic cells.

Secondary modifiers drive the expansion of LOY

Age is the strongest known risk factor for mLOY. While LOY is continuously and frequently acquired throughout life [12], mLOY prevalence increases rapidly in men beyond age 50 [30]. Although there is a pronounced intra-individual variation of changes in the frequency of LOY within Swedish and Danish men aged from 70 to 93, an increase with time is the most prominent longitudinal pattern of mLOY [31, 32].

Other non-modifiable risk factors for mLOY are germline mutations. To date, genome-wide association studies (GWAS) have identified around 358 loci in autosome and X contributing to mLOY risk, highlighting the polygenic architecture of mLOY (Fig. 3). Aligning with the disproportionate distributions of mLOY prevalence across racial and ethnic populations, the genetic risk loci also show specificities across them [33–35]. The majority of GWAS have primarily focused on subjects of European ancestry, the inclusion of other ethnic groups in GWAS is rising (Fig. 3). On pathway levels, GWAS hits have implicated in cell proliferation and survival (particularly in the context of cancer progression), cell cycle regulation, DNA damage response, blood cell differentiation, xenobiolic metabolism, somatic drivers of tumor growth, and targets of cancer therapy [33]. In this sense, it is interesting to decipher whether individuals with genetic predispositions to impaired DNA repair and increased tendency to develop malignancies (such as the so called chromosome instability syndromes [36]) have incremented risk for mLOY.

Fig. 3.

Number of mLOY-associated genetic loci identified by GWAS as a function of cohort size. This illustration shows that when sample size above a threshold, the rate of locus discovery accelerates exponentially. One of the first GWAS to access the genetic risk of mLOY was conducted in 13,789 men from three USA prospective cohorts and identified a locus at TCL1A was associated with the susceptibility for mLOY [37]. In 2017, a GWAS in 85,542 men of UKB replicated the reported genetic risk locus in TCL1A, and identified 18 novel loci [34]. In 2019, another GWAS performed in 205,011 participants in UKB, replicated all 19 known risk loci, and identified 137 novel loci [30]. Meanwhile, a GWAS study in 95,380 Japanese men from the cohort of Biobank Japan identified 50 independent genetic markers in 46 loci significantly associated with mLOY, 35 of which were novel [35]. A recent preprint work performed GWAS in 544,112 male participants of European, African, and Hispanic/Lantino ancestries in the Million Veteran Program. From this study, 323 genome-wide significant loci, 167 of which were novel, have been reported [33]. To avoid spurious associations rising from different allele frequencies among human ethnic populations, most GWAS of mLOY have considered subpopulations separately. Aligning with that the prevalence of mLOY is disproportionately distributed across racial and ethnic populations, genetic risk loci also show specificities across them

It seems that genetics does not explain all predisposition variance in mLOY. The epigenome-wide methylation analyses in 569 whole-blood samples identified 36 differentially methylated sites associated with mLOY and some of these methylation variable sites locate in genes involved in cell cycle regulation and chromosome segregation [34]. These data suggested that epigenetic mechanism may associate with mLOY. Epigenome-wide association studies aiming to comprehensively evaluate DNA methylation positions associated with mLOY are required.

In addition, mLOY is enriched in individuals challenged by environmental stressors and unhealthy lifestyles irrespective of age (reviewed in [38]). One possible scenario is that germline variants and acquired epigenomic alterations would determine the acquisition and expansion of LOY to a larger extent under aging or stressed conditions when intrinsic aging microenvironments or extrinsic environmental insults play a preponderant role.

Burden of LOY varies across tissues

In peripheral blood, LOY disproportionally affects certain cell types, with monocytes, natural killer cells, and regulatory T (Treg) cells being mostly affected [30, 39, 40]. While mLOY studies were pioneered in bone marrow-derived cell lineages, mLOY was also detected in multiple solid organs, such as the brain [41], kidney [42], and bladder [43]. Since the rise and expansion of LOY cells heavily depend on cell proliferation, it can be expected that, depending on cell proliferation rates and the timing of appearance of LOY mutation, the size of LOY clones varies profoundly among different tissue and cell types. In the human brain, aging-related LOY is frequently observed in microglia (proliferative), but rarely observed in neurons (terminally differentiated) [41].

In addition, clonal expansion of LOY cells may also be influenced by tissue intrinsic properties, such as cell–cell competition, space availability, and tissue stiffness. Comparisons across many different tissue types are essential for investigating tissue-specific fates of LOY cells, including LOY acquisition and the survival and proliferation of LOY cells. Mechanisms that lead to the rise of LOY cells and their clonal outgrowth in solid tissues are yet to be determined. By sampling blood cells and solid tissues from same individuals, it was found that mLOY occurs less frequently in the brain than in the blood [44]. Lower burden of LOY in solid tissues may either result from tissue-specific selection pressure, or from the spatial or physical restraints imposed by the tissue architecture. Thus, animal models are needed to study age-associated burden of LOY in different tissues and the regulatory mechanisms of clonal expansion of LOY cells throughout a lifespan.

Molecular alterations in LOY cells

Emerging multi-omics approaches that combine mutational, epigenetic, transcriptional, and biochemical profiling of LOY cells isolated from individuals with mLOY at a single-cell level have lend insight into molecular alterations upon LOY.

LOY is associated with genomic alterations

It has found that 9 of 12 LOY-positive men also bared CHIP and the number and variant allele fraction of affected SNVs were positively associated with cellular fraction of LOY in these men [45]. LOY is highly significantly associated with mutation in CHIP-driver genes TET2, DNMT3A, PPM1D, or ZNF318 in 5039 male Icelanders [46]. In addition, a study in 222,835 men from UKB found that CHIP is significantly associated with mLOY in ≥ 10% cells [47]. Such association is strengthened with the increase of LOY clone size and CHIP-driver mutations in TET2, TP53, and CBL is enriched in mLOY at clonal fractions > 30% [47]. In a study comprising 73 LOY patients with lymphoid or plasma cell malignancies, the frequency of a co-occurrence between mLOY and CHIP-driver mutations in blood is also associated with LOY burden [48]. In line with this, large-scale autosomal CAs occur more frequently in men with mLOY [37]. A comprehensive study of a 71-year-old male donor revealed that LOY HSCs are enabled to exit quiescence and potentially more permissive to acquiring 17p deletion [49].

Co-occurrence between LOY and other genomic alterations has also been observed in non-hematopoietic tissues and tumors. In the human kidney, LOY proximal tubule cells are more likely to harbor additional gains on chromosome 7, 10, and X [42]. Similarly, LOY, along with other CAs, could be observed in many tumor types [50]. Among 15 profiled LOY neuroblastoma, all of them have chromosome 17 aberrations and 11 of them have chromosome 11 aberrations (dominated by 11q deletion) [51]. LOY tumors display unique mutational profile and most significantly associate with driver mutations in TP53 [52, 53].

Why CHIP or autosomal CAs co-occur more frequently with LOY? The first explanation is that these mutations synergistically promote the clonal advantage and/or influence HSC differentiation. This model is compelling because, to compensate harmful effects of aneuploidy, cells would adapt to a new genomic steady state by inducing mutations ranging from DNA levels to chromosome levels [54]. The second explanation is that these CH events have shared genetic origins. Many genetic variants have been linked to the development of LOY, CHIP, and autosomal CAs [55] are shared, such as those affecting TCL1A, CHEK2, ATM, and TERT that primarily involved in genome stability maintenance [55]. Thus, genetic variants predisposing to LOY may facilitate the acquisition and outgrowth of clones bearing CHIP or CAs. The third explanation is that LOY and autosomal CAs are derived by CHIP. For example, chromosome numerical and structural aneuploidies are tolerated and propagated in p53-deficient cells [56]. The fourth explanation is that LOY is able to induce GIN. Pioneering technologies have dramatically improved the generation of LOY cells. In mouse hematopoietic stem and progenitor cells (HSPCs) and tumor cells, LOY clones induced by genome editing have increased level of DSBs [5]. This observation, although not definitive, suggests that acquisition of LOY mutation might bolster cells to obtain new mutations. Further work is required to determine if LOY can causally link to CHIP or autosomal CAs in immune cells.

LOY is associated with epigenomic alterations

A considerable body of evidence suggests Y has a major effect on epigenetics of Y and other chromosomes. The assembly length of Y is negatively associated with global levels of DNA methylation at CpG sites, both genome wide and Y wide [10]. The Y-linked KDM5D, which encodes a lysine demethylase, may be one candidate gene responsible for Y’s effect on autosomal DNA methylation [57]. In KRAS-mutated colon cancers, KRAS activates STAT4 to upregulate KDM5D, which represses expression of regulators involved in epithelial cell tight junction and tumor immune recognition via histone demethylation and deacetylation [58]. In addition, a total of 194 Y genetic variants have found to impact DNA methylation in human Y, and DNA methylation is varied in 2861 segments across Y and 21% of them are associated with Y haplogroups [10].

Growing evidence support a role of LOY in epigenetic alterations. The transposable elements of constitutive LOY mice are globally demethylated, as compared with those of the wildtype counterparts [57]. A recent work found that DNA methylations of numerous genes already implicated in AD pathogenesis are influenced by LOY [59], and differential DNA methylations occurring in gene regulatory regions are predominantly accompanied by downregulation of affected genes [60]. UTY seems to be one major gene accounting for LOY-induced epigenetic alterations in both monocytes and macrophages [61]. Since LOY impacts genome-wide DNA methylation and mLOY prevalence is age-dependent, further mining the epigenetic data in large cohorts will facilitate the discovery of whether mLOY is associated with epigenetic age acceleration, as what has been observed in age-related CH [62]. In addition, as human Y is enriched for non-coding RNAs that remain largely insufficiently studied [63], future studies aiming at characterizing non-coding RNAs alterations in LOY cells would add a new layer of complexity in LOY-induced epigenomic alterations.

LOY induces transcriptomic alterations

Several lines of evidence from single-cell RNA sequencing indicate that LOY broadly regulates the expression of autosomal and X-linked genes in cell-type specific ways. About 500 autosomal genes, many of these involved in physiological immune surveillance, are altered by LOY in leukocytes from AD patients [39]. Single-cell transcriptomics reveals that AD-specific immune activation is dominated in LOY immune cells from AD patients [59]. In these LOY cells, several genes involved in leukocyte differentiation exhibit differential methylation in promoters and dysregulation expression [60]. LOY microglia have altered expression of 172 autosomal genes, 3 X-linked genes, and 10 pseudoautosomal genes with diverse functions [41]. Across cell types of human kidney, LOY is associated with decreased expression of DNA repair and cell cycle genes and kidney proximal tubule cells with LOY are more likely to express genes involved in apoptosis, cell transport, and differentiation [42]. Gene ontology analysis of differentially expressed genes between HSCs with and without LOY identifies altered pathways linked to HSCs quiescence [49]. Unsurprisingly, extent of transcriptomic alterations associated with mLOY depends on cellular fraction of LOY [64].

Transcriptomic studies in induced and spontaneous-occurring LOY cells revealed that LOY drives transcriptional dysregulation. Celli et al. [65] used the CRISPR/Cas9 technology to induce LOY in human non-transformed ARPE-19 cells and found that LOY influences the transcription of genes involved in cell migration regulation, angiogenesis, and immune response. Sano et al. [4] sequenced the LOY immune cells generated by CRISPR/Cas9 technology and revealed that LOY activates a signaling network to promote macrophage polarization toward a fibrotic phenotype. Human-induced pluripotent stem cells from LOY fibroblasts show disrupted expression in genes involved in cholesterol metabolism, neural crest cell signaling, and translation [66]. Abdel-Hafiz et al. [67] sequenced the entire transcriptome in MB49 tumors with spontaneous-occurring LOY grown in wildtype mice and revealed that LOY tumors increase the expression of genes associated with suppressed immune responses compared with non-LOY tumors. These data demonstrate a causal link between LOY and genome-wide transcriptomic alterations. Indeed, several Y-linked genes, such as ZFY, have significant regulatory effects to drive cell-type-dependent autosomal and X transcriptional responses, both in vitro and in vivo [68].

mLOY is associated with proteomic and metabolomic changes

In area of mLOY, proteomic and metabolomic studies currently lag behind transcriptomic studies. Two recent studies accessed clinical serum biomarkers in participants from UKB, and LOY is associated with changes in levels of multiple serum biomarkers, namely, lipid, sex hormones (testosterone), sugar, vitamins, and minerals [47, 69]. The same studies found that mLOY participants had altered serum levels of several proteins. Particularly, LOY was strongly associated with elevated levels of sex hormone binding globulin. However, the causal relationship between mLOY and proteomic and metabolomic changes in serum is still being unveiled. As the technologies advance, single-cell-based proteomic and metabolomic studies of LOY cells will help to draw a broader picture of molecular alterations in LOY cells.

Cellular and subclinical alterations after LOY

As mLOY burden varies based on cell types, the associated molecular and cellular alterations are expected to exhibit cell type specificity. In this section, we draw on insights gained from human single-cell-based studies and the continuing researches in experimental models that have revealed mLOY drives diverse cellular and subclinical phenotype alterations.

LOY drives cell transformation, immune evasion, and metastasis

Cancer is the prime example to study the precise mechanisms by which LOY drive cellular remodeling. Zhang et al. [5] have constructed the first mouse model of hematopoietic mLOY (Fig. 4A). Given that mLOY is frequently observed in AML1-ETO⁺ AML, the authors aimed to dissect the role of mLOY in leukemogenesis. In a key experiment, LOY HSCs in the background of Trp53^–/– were transduced with AML-ETO and the resulting HSCs were then transplanted to sub-lethally irradiated wildtype female mice. Notably, the results showed that LOY-AML1-ETO mice have a significantly shorter latency to develop AML than AML1-ETO mice. In a competitive clongenic bone marrow transplant assay, irradiated mCD45.2 mice were reconstituted with LOY HSCs from mCD45.1 mice and the authors discovered an age-dependent clonal expansion of LOY (Fig. 4B). In this respect, the most straightforward mechanism underlying LOY-mediated AML acceleration would be that the amplified LOY clones possess an increased chance of transformation into AML.

Fig. 4.

Summary of detailed methodology for modeling mLOY in hematopoietic system of mice. Experimental conditions to model mLOY in studies of Zhang et al. [5] (A, B) and Sano et al. [4] (C). One fundamental difference between these two studies is that the recipient mice were males in study of Sano et al. but were females in study of Zhang et al. While Zhang’s model has an advantage to rapid calculate the contribution of LOY cells to leukemogenesis, it has a drawback because this exclusively happens in mLOY men. Instead, this condition happens when female recipient receiving allogeneic HSC transplantation from male donors with LOY. Abbreviations: HSCs, hematopoietic stem cells; HSPCs: hematopoietic stem and progenitor cells; LOY, loss of Y chromosome; sgRNA, small guider RNA

In rare cancer uveal melanoma, clonal LOY is strongly associated with tumor metastasis and heterogenicity, suggesting LOY is a potential driver event [53]. In line with this view, LOY induced by CRISPR/Ca9 promotes cell migration and invasion of non-transformed human ARPE-19 cell line [65]. Abdel-Hafiz et al. [67] found that LOY MB49 cell lines showed a twofold increase of proliferation than non-LOY lines when being injected subcutaneously into immune-competent mice. Furthermore, they found that LOY tumors create an immunosuppressive tumor microenvironment via promoting CD8⁺ T cell exhaustion. Based on these data, they further revealed that CD8⁺ T cells from LOY tumors are more responsive to anti-PD1 treatment than those from non-LOY tumors. These results demonstrate that LOY alters immune cell composition in the tumor microenvironment. In line with this, a recent study uncovers that, by diminishing the expression of several autosomal genes that encode cancer/testis antigens, LOY alters cell composition of tumor immune microenvironment and allows LOY cancer cells to avoid immune attack, giving them an advantage over non-LOY tumors [29].

Hematopoietic LOY induces cardiac fibrosis

Sano et al. [4] developed a mLOY mouse model, in which LOY is observed in 65% of blood cells and this level of mosaicism can be maintained for 12 months (Fig. 4C). mLOY mice have shortened lifespan, accelerated tissue fibrosis, and impaired short-term working memory. Furthermore, they found that young LOY mice after transverse aortic constriction have greater interstitial and perivascular fibrosis in heart and this pathological cardiac phenotype can largely be explained by alterations in LOY myeloid cells. RNA sequencing of LOY immune cells reveals that LOY activates a signaling network to promote macrophage polarization toward a fibrotic phenotype. Pro-fibrotic signaling from LOY macrophage is the primary culprit underlying experimental cardiac dysfunction, as its neutralization by transforming growth factor (TGF) β1 antibody reverses the cardiac pathology.

In line with this, a recent study found that mLOY is linked to a higher three-year mortality in patients after transcatheter aortic valve replacement, even confounding factors are adjusted [70]. Single-cell RNA sequencing revealed that LOY specifically influences monocytes and natural killer cells. In concordance with observations in mouse [4], LOY monocytes display a pronounced pro-fibrotic gene signature characterized by the upregulation of TGFβ-associated signaling. Together, these results indicate that pathological mechanisms of mLOY in driving cardiovascular diseases (CVD) maybe conserved in mice and human.

Hematopoietic mLOY is associated with taupathology

AD is the most widely form of dementia characterized by plaques comprising of misfolded amyloid-β (Aβ) and neurifibrillary tangles comprising of hyperphosphorylated tau in the brain. Compared to other neurodegenerative diseases, LOY is significantly enriched in AD subjects [41]. A seminal study has demonstrated that Sweden men with mLOY in blood are 6.8 times more likely for AD diagnosis [71]. Such association is validated, although to a less extent, in US-based cohorts [72]. Recently, a study in two geographically distinct cohorts replicated the association between mLOY and AD [73]. Interestingly, individuals with higher proportions of LOY in blood have increased levels of AD pathogenic biomarkers in cerebrospinal fluid, including total tau and phosphorylated tau-181, but not Aβ₄₂ [73].

Regarding hematopoietic system, patients with AD are primarily affected with LOY in natural killer cells [39], a type of immune cell important for controlling viral infection and their accumulation in aging brain impairs neurogenesis and cognition [74]. Besides, mLOY has been detected in brain cells of patients with AD. Given the diversity of cell populations in the human brain, are there particular brain cell atlases that are predisposition to LOY? Indeed, age-dependent mLOY is highly enriched in microglia, mildly observed in astrocytes and oligodendrocytes, but rare in neurons [41]. Since microglia are long-lived resident macrophage population arisen from a distinct lineage compared to other brain cells [75], enrichment of LOY in microglia suggests LOY may confer a selective advantage in this cell type and cause susceptibility to death in other brain cell types.

Hematopoietic mLOY is associated with weakened antiviral immune responses

A recent work in Spanish cohort has unrevealed that COVID-19 patients with mLOY in blood had a 40% increase in risk of lethality within 90 days after infection [76]. Another work analyzed different blood cells of 139 critically ill male COVID-19 patients in Sweden and found myeloid lineage cells (such as low-density neutrophils and monocytes) that are crucial for development of severe COVID-19 exhibit the highest levels of clonal LOY [77]. These data suggests that HSCs with LOY may differentiate in a myeloid-biased way. Support for this notion is building: in mouse HSCs, LOY lead to an increased production of mature myeloid cells, specifically macrophages, in the cardiac tissues [4].

Although these findings support that mLOY acts as a patient-intrinsic factor in modulating COVID-19 severity, no unified mechanism of susceptibility has yet been proposed. It has shown that 13 genes involved in response to SARS-CoV-2 infection are downregulated in mLOY individuals [76]. Similarly, single-cell transcriptomics in sorted CD14⁺ monocytes from COVID-19 patients and controls has shown that the transcription of genes important for antigen presentation were pervasively downregulated [77]. Thus, it is tempting to speculate that mLOY would weaken the antiviral immune responses in men, thereby strongly exacerbating sex-based outcomes after SARS-CoV-2 infection. Interestingly, recovery from COVID-19 is associated with significant decreases in LOY prevalence in the blood [77], raising an important question of whether SARS-CoV-2 infection facilitates the acquisition and expansion of LOY.

Paracrine effect of LOY cells to non-LOY cells

In tumor cells, contact between cells carrying different genetic alterations would result in the emergence of either cell cooperation or cell competition [78]. Behavior consequence of interactions between cells carrying divergent functional variants depends on the properties of each genetic variant, population size, and tissue compartmentalization [78]. Recently, it has found that supernatants from cultured DNMT3A mutant macrophages activate the pro-inflammatory genes in wildtype macrophages, which in turn induce pro-hypertrophic effect to cardiomyocytes [79]. In addition, supernatants from cultured DNMT3A mutant macrophages activate human cardiac fibroblasts by stimulating wildtype CD4⁺ T cells. These results suggest a novel paracrine effect of DNMT3A mutant to wildtype cells and paracrine effect maybe mediated by metabolites or pro-inflammatory factors produced by DNMT3A mutant cells.

Whether there are functional interactions or paracrine signaling pathways between LOY and non-LOY cells is current unknown. Building on new techniques developed by Abplanalp et al. [79], we have a robust tool to elucidate the paracrine effect of LOY cells to non-LOY cells. A recent work analyzed mLOY in tumor-infiltrating immune cell atlas and showed that Treg cells have the highest rate of LOY among the subpopulations of T lymphocytes in primary colorectal cancers and liver metastases [80]. This study found that LOY Treg cells upregulate the expression of immunosuppressive genes in colorectal cancer patients [80]. Thus, LOY Treg cells infiltrated in tumor microenvironment provide a unique scenario to study the paracrine effect of LOY Treg cells to non-LOY cancer cells.

Clinical phenotypes associated with mLOY

The clinical consequences of mLOY depend on the combined consequence of cellular alterations. In this section, we discuss the clinical phenotypes associated with mLOY in three contexts: hematopoietic tissue, solid healthy tissues, and solid tumors.

mLOY in hematopoietic tissue

In 2014, a seminar study found that, in healthy cohorts, detection of mLOY in blood is associated with increased risk of non-hematological cancers and all-cause mortality [3]. Since then, a broader assessment of blood mLOY in population-scale cohorts has confirmed these associations and linked mLOY to a heightened risk for additional age-associated pathologies, such as solid cancers, AD, cardiovascular events, and age-related macular degeneration (reviewed in [17]). The finding that blood mLOY is linked to a wide range of non-hematologic disorders seems unsurprising, given that, through their potent circulating property, LOY blood cells are in constant contact with other tissues and capable of altering tissue and organ function throughout the body. The list of mLOY-associated disorders is still expanding. For example, mLOY is recently found to be associated with poor outcome after ischemic stroke in patient not receiving recanalization therapy [81], the age of onset and duration of schizophrenia [82], and increased risks of all lung diseases [83]. A recent finding that mLOY is associated with the occurrence and increased mortality of wildtype transthyretin cardiac amyloidosis [84] suggests mLOY may exert pathological role in disorders of proteostatic dysfunction other than AD.

It has shown that, in some cohorts, individuals carrying mLOY are only significantly associated with increased myeloid malignancy risk when also harboring CHIP or autosomal CAs [85]. This result suggests that there is a synergistic effect of these mutations on disease risk, but thus far experimental evidence for the causal effect is still lacking. Deciphering whether and how the co-expansion of CHIP or CAs with LOY confers higher risks of disease incidence or worse outcomes than either one alone is needed for a better understanding pathological consequence of mLOY.

Despite mLOY is a trait specific to men, the polygenic risk score (PRS) for mLOY is also linked to disease risk in women. A study imputed PRS data comprised of 156 lead variants linked to mLOY in men and found a significant association of PRS with breast cancer and later age at menopause [30]. Similarly, increasing PRS for mLOY has found to be associated with poor functional outcome after ischemic stroke in women [86]. How to explain these observed associations between PRS for mLOY and disease risk in men? One hypothetical scenario is that these associations may reflect confounding by shared genetic predisposition. Another hypothetical scenario is that mLOY is not mechanistically linked to pathogenesis in men, instead both of them are potentially both outcomes of an inherited predisposition for GIN.

mLOY in solid tumors

While mLOY in blood is associated with the risk of many cancer types, cytogenetic studies have long revealed that LOY could be observed in many solid tumor types and LOY is part of a complex karyotype in tumors [50]. A recent study carried out a comprehensive analysis of LOY across over 5000 male tumors from The Cancer Genome Atlas [53]. Similarly, another study analyzed the genomic and transcriptional data of 13 tumor types from 2375 patients [52]. These two studies consistently provided several novel insights into our understanding of mLOY in tumors. (i) LOY occurs frequently in primary cancers, with 30% male tumors harboring either complete or relative LOY. (ii) LOY frequency varies substantially across tumor types, ranging from nearly 80% in renal papillary cancer to < 2% in pheochromocytoma and paraganglioma and thymoma, suggesting the differential dependence of tumors on LOY. (iii) LOY is more commonly observed in tumors with GIN, including SNVs and CAs. (iv) Tumor cells display wide distributions of LOY cell fraction, ranging from < 0.4 in prostate adenocarcinoma to 0.9 in uveal melanoma, suggesting tumor LOY can arise via both clonal and subclonal processes. (v) LOY tumors display unique mutational profile and LOY is most significantly associated with driver mutations in TP53. (vi) Individuals with LOY tumors have a shorter overall survival compared with non-LOY tumors. (vii) LOY is a potential driver in rare cancer uveal melanoma, where clonal LOY is strongly associated with tumor metastasis and heterogenicity. (viii) Male-dominated tumor types have a higher frequency of LOY, suggesting LOY might contribute to male-biased cancer development.

mLOY in solid tissues

Compared to mLOY in hematopoietic tissue, relatively little information is available so far on mLOY burden in solid tissues due to difficulties of sampling. mLOY has been found in human dorsolateral prefrontal cortex and cerebellum and is functionally linked to neuropathological and clinical characteristics of AD and cognitive aging [44]. Regarding brain cell types, microglia is the mostly affected by mLOY [41]. Moreover, LOY frequency was elevated in AD microglia compared with controls [41 Besides, a survey of genomic alterations via whole genome sequencing showed that LOY is detected in a high cell fraction (~25%) in the postmortem brains of patients with autism spectrum disorder, pointing to the relevance of mLOY for the autism phenotype [87]. mLOY is also detected in the kidney and is associated epithelial injury in chronic kidney disease [42]. In the kidney detected with mLOY, cortical epithelial cell types are frequently affected [42]. In addition, mLOY is detected in histologically normal urothelium from bladder urothelial carcinoma patients and LOY burden normal urothelium is slightly lower than this of urothelial carcinoma [43]. Coupling with earlier finding that LOY is frequent early event in bladder urothelial carcinoma [88]. LOY expanding in normal urothelium may endow the clones with tumorigenic properties.

Together, mLOY in solid tissues is also associated with worsened pathological outcomes in the affected tissues. While we focused on blood-based studies, studying mLOY in other easily accessible non-blood system, such as skin and sperm, in large-scale cohorts, would also be highly informative for understanding the clinical potential of mLOY.

Implications of the UYC model for therapeutic interventions

The UYC model provides a new paradigm for studying the biological and clinical significance of mLOY. Considering clinical importance of mLOY, a broader examination of whether mLOY could be a disease-halting intervention target is necessary [38]. Particularly, since mLOY has a direct effect in cancer progression [67], targeting LOY is an emerging therapeutic strategy for patients with mLOY tumor. According to the UYC model, various conceptual opportunities for designing such interventions exist.

Preventing the arise and expansion of LOY

In mLOY carriers, size of LOY clones correlates with disease risk [48]. Thus, inhibiting clonal expansion of LOY is a strategy for preventing the adverse disease outcomes associated with mLOY. The prevalence of mLOY increases slowly over decades in men aged below 50, but increases dramatically beyond the age of 50 [30]. The trigger for shifting the developmental trajectory of LOY cells is as-yet unknown, the most potential candidate is HSCs (including HSPCs) aging [89], given that blood mLOY is driven by LOY mutation in HSCs. Although chronological aging (measured in years) is un-modifiable, biological aging (measured with molecular and cellular biomarkers) occurs on different scale and is therefore modifiable. In recent year, we have witnessed a growing number of pharmacological agents that are effective at combating aging and aging-related diseases [90]. It will be interesting to understand whether geroprotectors might be used to decelerate clonal expansion of mLOY.

Although human cohort studies have suggested several environmental stressors and unhealthy lifestyle are associated with clonal expansion of mLOY, greater mechanistic studies in animal models are needed to adequately test the causal relationship. If environmental stressors and/or modifiable lifestyle are causally linked to mLOY, the incidence of mLOY should be inhibited by a general improvement in environment and lifestyle health. Besides, mLOY is associated with changes in levels of multiple serum biomarkers, namely, lipid, sex hormones, sugar, and vitamins [47, 69]. Thus, it is interesting to determine whether there is extensive crosstalk between age-related serum microenvironments and clonal expansion of LOY, which would suggest novel intervention avenues for preventing mLOY. Since serum biomarkers are dramatically affected by lifestyle, an in-depth understanding of the extent to which healthy diets, dietary restriction, and regular exercise can be optimized to decelerate mLOY is needed.

Principally, the clonal expansion of LOY can be blocked by pharmacologically targeting genetic loci that increase the fitness of LOY HSCs. TCL1A could be a prime promising drug target for preventing mLOY. While aberrant activation of TCL1A drives the clonal expansion of LOY, probably via altering cell cycle and DNA repair pathways [91], a commonly inherited variant (the T allele of rs2887399) restricts chromatin accessibility of TCL1A promoter, resulting in decreased TCL1A expression and abrogated clonal expansion of LOY cells [92]. Thus, a logical approach to mitigating mLOY is to seek ways to downregulate the expression or activity of TCL1A.

Another challenge is to identify time windows when interventions have the greatest effect on the clonal expansion of mLOY. Since the increment of mLOY is age-dependent, early initiation of interventions might lead to better results. Although LOY clones dominated in old age are predicated to originate before the age of 40 [13], 50 years of age is a threshold for mLOY prevalence increment in UKB [30], These data indicate that the clonal expansion of LOY occurs slowly over decades. Whether setting the intervention threshold to ~ 50 years could decelerate mLOY development in genetically and/or environmental susceptible individuals deserves further study. Before this, long-term longitudinal studies in large-scale cohorts are needed to evaluate whether mLOY PRS is predictive in identifying individuals at risk of developing mLOY, and animal studies are required to clarify to what degree age-related increase of mLOY burden are due to environmental factors.

Eliminating the already-formed LOY clones

The second strategy aims to eliminate the already-formed LOY clones by targeted drug development. In some aged Swedish and Danish men, the longitudinal frequency of mLOY remains unchanged or decreases with time [31], and clonal expansion of mLOY in later life could be inhibited or reversed. Since LOY leads to cellular stresses, such as deficient DNA damage response [65], they endow affected cells with unique vulnerabilities. Compounds that selectively cause lethality in LOY cells, either by exaggerating the adverse stresses of LOY or by interfering with pathways that are essential for cell survival, are yet to be developed. As LOY is a specific form of aneuploidy, exploring whether LOY cells are sensitive to compounds that have potential to selectively targeting aneuploidy [93, 94] is warranted. A growing number of studies have identified drugs that could target specific recurrent segmental monosomy, such as loss of 8p and 2q [95], providing exciting new opportunities for identifying compounds that can specifically target LOY cells.

Targeting the consequences of mLOY

The third strategy aims to target the pathological consequences of mLOY. While mLOY induced a relative increase in macrophages expressing pro-fibrotic genes, the antibody neutralizing TGFβ1 signaling reversed the cardiac dysfunction observed in mLOY mice [4], demonstrating that, at least in preclinical mouse models, blocking the downstream effects of mLOY could serve as an effective approach to inhibit the pathogenic effect of mLOY. LOY cancer cells alter T cell function and promote T cell exhaustion, making LOY tumors more susceptible to anti-PD-1 treatment [67]. In addition, a growing number of studies has revealed that deletion of one key MSY gene is sufficient to recapitulate pro-pathogenic effects of mLOY in AML (Kdm5d) [5], cardiac dysfunction (Uty) [61], or bladder cancer (Uty or Kdm5d) [67]. These findings provide critical clinical opportunities for targeting mLOY via increasing the expression of MSY genes in non-LOY cells, thereby compensating the loss of MSY genes in LOY cells.

Interestingly, a recent prospective study in Chinese adults found that while mosaic autosomal loss in blood is associated with increased risk of bladder cancer, this is only evident among participants with low physical activity [96]. This result suggests that exercise ameliorates disease risk associated with mosaic autosomal loss, even when it has no profound effect on mosaic autosomal loss development. It will be interesting to decipher whether pathological consequences of mLOY could be ameliorated by lifestyle intervention. Unexpectedly, data from UKB shows that males with moderate to high physical activity have elevated risks of mLOY [97]. Whether this result is specific to UKB participates or is generalizable to other cohorts should be tested by future studies.

Conclusions and future perspectives

Although mLOY has been reported for over six decades, rapidly developing technologies, long mystery of human Y, and incompletely understood sexual dimorphism in human aging, disease susceptibility, and lifespan have kept this field both active and quickly evolving in the last decade. The pervasive prevalence of mLOY in aging men suggests a potentially pivotal role in aging and age-related diseases. However, it is challenging to determine whether there is a causal relationship in humans. Emerging multi-omics approaches that combine gene expression, epigenetic profiling, and mutational profiling of human LOY cell populations at a single-cell level have found that LOY correlates with genomic, epigenomic, and transcriptomic alterations, but whether these are causal or correlative is not clear. Fortunately, contemporary work in preclinical models has begun to investigate the causality and novel pathological mechanisms of mLOY.

Despite these, understanding why, how, where, and when does mLOY cause a myriad of age-related diseases, is still far from complete. This review has attempted to collate the current understanding of mLOY into the UYC model that can provide a general framework to deep our understanding of mLOY. We extrapolate the Liebigs’ law of minimum to propose that the level of genome stability of male cells, and the overall health of organism they consisted depends on stability of the most unstable chromosome—Y chromosome (Fig. 5). This is the fundamental concept driving the UYC model. We believe that the UYC model will provide a new paradigm for studying the biological and clinical significance of mLOY and make a number of testable predictions that will stimulate further mechanistic work in this direction and help to support the theory. Undoubtedly, insights obtained from the UYC model represent only a portion of the broader landscape of mLOY yet to be understood.

Fig. 5.

Extrapolating the Liebigs’ law of minimum to understand the unstable Y cascade (UYC) model. The law of minimum, firstly proposed by Justus von Liebig in 1840, states that “the yield is proportional to the amount of the most limiting nutrient, whichever nutrient that may be.” This concept can be broadened to understand the UYC model. If we redefine “yield” as the level of genome stability and “most limiting nutrient” as most unstable chromosome, then we can say that the level of genome stability of male cells, and the overall health of organism they consisted, depend on the stability of the most unstable chromosome—Y chromosome. Loss of the unstable Y disrupts the balanced forces in maintenance of genome stability; leads to the loss of genetic, epigenetic, and transcriptomic information; and thereafter induces a cascade of tissue-specific cellular alterations that contribute to aging and onset and progression of age-related diseases. In other words, healthy aging of males could be rooted in the Y chromosome

We anticipate at least two major areas of development. First, a noteworthy aspect is to clarify if and to what extent mLOY does in fact drive aging. Although the prevalence of mLOY in human populations is age-dependent, the direct link between mLOY and human aging remains unclear. Since mLOY in blood is associated with decreased lifespan in human [3] and mice [4], it is plausible that mLOY, along with its downstream effects proposed in UYC model, might drive organism aging. To test it, targeted and comprehensive studies in animal models are required. Since tissue aging and disease-associated pathologies are broadly intertwined, examining the pathological mechanisms of mLOY in aging tissues may provide invaluable insights into the potent aging-driving effects of mLOY. If this is true, mLOY is a novel underlying cause of sex differences in human aging and lifespan [6]. Second, novel studies are needed to better understand the causal role of mLOY in age-related diseases other than cancer and CVD. A prime candidate is AD [98]. Association between mLOY and AD risk has replicated in several independent cohorts [71–73]. The preliminary result has shown that mLOY mice exhibit short-term working memory [4], suggesting it is valuable to study the role of mLOY in better models of AD. In designing AD models, we need to consider incorporating hematopoietic mLOY or microglia-specific mLOY into the mice with human AD causative genes or risk genes being knocked in, such as 5xFAD mice and APOE4 knock-in mice.

It is worth to note that mLOY in blood is not always associated with adverse clinical outcomes. For example, mLOY is not significantly associated with COVID-19 diagnosis or hospitalization in UKB [99]. Several independent studies have suggested that mLOY is not associated with the risk of all-cause death [100], the overall or specific cancer risk [37], lung cancer [101], AD, and CVD [102]. While mosaicism burden is a known factor influencing the phenotype of mLOY, mLOY can be [103] or be not [101] associated with lung cancer in two different cohorts using the same threshold. More remarkably, mLOY is associated with clinal benefits. For example, mLOY is associated with reduced risk of retinal disease and with a 10–35% reduction in the risk of some metabolic diseases like type II diabetes in UKB [102]. The spontaneous mLOY, not this related to smoking, is associated with reduced lung cancer risk in non-smoking Chinese [104]. While mLOY is positively associated with late-onset rheumatoid arthritis, it is negatively associated with young-onset rheumatoid arthritis in Japanese [105]. Of interest, mLOY is inversely associated with mortality in Danish centenarians [32]. These findings demonstrate that mLOY-related disease risk in different cohorts probably depends on the overall health status and the ethnic background. Beneficial outcomes associated with mLOY are relatively unexplored, meriting additional studies of underlying biological mechanisms.

Abbreviations

Aβ

Amyloid-β

AD

Alzheimer’s disease

AML

Acute myeloid leukemia

CAs

Chromosomal alterations

CH

Clonal hematopoiesis

CHIP

Clonal hematopoiesis of indeterminate potential

CVD

Cardiovascular disease

DSBs

Double-strand breaks

GIN

Genomic instability

GWAS

Genome-wide association study

HSCs

Hematopoietic stem cells

HSPCs

Hematopoietic stem and progenitor cells

LOX

Loss of X chromosome

LOY

Loss of Y chromosome

mLOY

Mosaic loss of Y chromosome

MSY

Male-specific region of Y

PARs

Pseudoautosomal regions

PRS

Polygenic risk score

SNVs

Single-nucleotide variants

TGF

Transforming growth factor

TME

Tumor microenvironment

Treg

Regulatory T

UKB

UK biobank

UYC

Unstable Y cascade

Author contribution

X. Guo had the idea for the article and constructed the figures. X. Guo and X. Dai performed the literature search and data analysis, and X. Guo and X. Dai drafted and/or critically revised the manuscript. All authors approved the final version of the manuscript and agreed on the order in which their names were listed in the manuscript.

Funding

X.G. is funded by the National Natural Science Foundation of China (No. 32260148, 31900410). X.D. is funded by the National Natural Science Foundation of China (No. 82302957), the Yunnan Fundamental Research Projects (No. 202401CF070054), and Biomedical Projects of Yunnan Key Science and Technology Program (202302AA310046).

Data availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Disclosure

Xihan Guo is assigned as corresponding author and acts on behalf of all co-authors and ensures that questions related to the accuracy or integrity of any part of the work are appropriately addressed.