Genetic models for lineage tracing in musculoskeletal development, injury, and healing

Abstract

Musculoskeletal development and later post-natal homeostasis are highly dynamic processes, marked by rapid structural and functional changes across very short periods of time. Adult anatomy and physiology are derived from pre-existing cellular and biochemical states. Consequently, these early developmental states guide and predict the future of the system as a whole. Tools have been developed to mark, trace, and follow specific cells and their progeny either from one developmental state to the next or between circumstances of health and disease. There are now many such technologies alongside a library of molecular markers which may be utilized in conjunction to allow for precise development of unique cell ‘lineages’. In this review, we first describe the development of the musculoskeletal system beginning as an embryonic germ layer and at each of the key developmental stages that follow. We then discuss these structures in the context of adult tissues during homeostasis, injury, and repair. Special focus is given in each of these sections to the key genes involved which may serve as markers of lineage or later in post-natal tissues. We then finish with a technical assessment of lineage tracing and the techniques and technologies currently used to mark cells, tissues, and structures within the musculoskeletal system.

1.0. Introduction:

Movement, structure, and support rely on a complex biomechanical network of static and dynamic tissues we define as the musculoskeletal system. Each component of the musculoskeletal system is highly specialized in form and function with unique approaches to homeostasis and repair. Each of these tissues in turn arose from a more primordial state, often developing via shared ancestral cells or tissues in a ‘lineage’ which can be traced across time. Adult form and function are derived from the interaction of these earlier states and, consequently, study of these lineages provides critical data to predict how adult tissues will respond to stress, injury, and disease. Because developmental states are typically transitory, the challenge of this study is accurate identification and tracking of each lineage across time. A range of tools have been developed to allow for precise ‘lineage tracing’ across animal models for study of musculoskeletal health and disease. In this review we will focus on those lineages which lead to the bones, muscles, cartilage, tendons, and ligaments necessary to facilitate structure and movement. We will explore the development of these structures, those markers which may be used to follow that development, and transitory patterns of expression which may be used to study key stages throughout.

2.0. Origins of the Musculoskeletal System:

Immediately after fertilization, the human ovum divides through several stages including the morula, blastocyst, and gastrula.^1,2 The gastrula is a tri-laminar structure from which we derive the three commonly classified germ layers of ectoderm, endoderm, and mesoderm. The outer layer or ectoderm will become the nervous system and the epidermis and is the source for the neural crest cells. Neural crest cells are a migratory cell population which contributes to the musculoskeletal system of the head and neck. The inner layer or endoderm gives rise to the internal epithelium including both digestive and respiratory systems and their associated organs. The intermediate layer or mesoderm gives rise to a vast majority of the body’s structures including the substructures of the musculoskeletal system, the dermis, and the cardiovascular system.³ The mesoderm classically defines the musculoskeletal system and can be marked at this stage by the transcription factor brachyury also known as T-box transcription factor T (TBXT). Brachyury is necessary for mesodermal differentiation and, consequently, marks the entirety of the mesodermal lineage (Table 1&2).^4–5

Table 1:

Sample Cre, CreERT, and Floxed Lines. Genes allowing for recognition and tracing of specific musculoskeletal lineage targets are represented as well as the class of modification (Cre, CreERT, Floxed). References provided are for manuscripts which first reference the generation of selected lines.

Gene	Target Cells	Mouse Lines	References
TBXT (Brachyury)	Mesoderm	Cre	Perantoni et al., 2005
		CreERT	Anderson et al., 2013
TBX6	Paraxial Mesoderm (Pre-Somitic)	CreERT	Peter Lopez et al., 2012
DLL1	Paraxial Mesoderm (Pre-Somitic)	Cre	Ji et al., 2006
PAX3	Paraxial Mesoderm (Pre-Somitic); Muscle	Cre	Lang et al., 2005
		CreERT	Southard et al., 2014
MEOX1	Paraxial Mesoderm (Somitic)	Cre	Jukkola et al., 2005
FOXF1	Lateral Plate Mesoderm	Flox	Ren et al., 2014
PRRX1	Lateral Plate Mesoderm; Pre-Chondral Mesenchyme	Cre	Logan et al., 2002
		CreERT	Kawanami et al., 2009
PAX7	Muscle Satellite Cells	Cre	Lepper et al., 2009
		CreERT	Lepper et al., 2012
		Flox	Lepper et al., 2009
MYF5	Adult Muscle, Muscle Satellite Cells (Proliferating)	Cre	Tallquist et al., 2000
		CreERT	Biressi et al., 2013
MYF6 (MRF4)	Adult Muscle, Muscle Satellite Cells (Differentiating)	Cre	Keller et al., 2004
		CreERT	Southard et al., 2014
MYOD1	Adult Muscle, Muscle Satellite Cells (Proliferating)	Cre	Haldar et al., 2014
		CreERT	Southard et al., 2014
MYOG1	Adult Muscle, Muscle Satellite Cells (Differentiating)	Cre	Li et al., 2005
		CreERT	Southard et al., 2014
MCK/CKM	Adult Muscle	Cre	Bruning et al., 1998
MYL 1/3 (MLC1/3)	Adult Muscle	Cre	Bothe et al., 2000
		CreERT	Southard et al., 2014
NOTCH1	Muscle Satellite Cells (Quiescent)	Flox	Yang et al., 2004
STAT3	Muscle Satellite Cells (Proliferating)	Flox	Moh et al., 2007
SCX	Tenocyte/Ligamentocyte/Enthesis	Cre	Sugimoto et al., 2013
		CreERT	Howell et al., 2017
MKX	Tenocyte/Ligamentocyte/Enthesis	Flox	Liu et al., 2010
GLI1	Tenocyte/Ligamentocyte Progenitors	CreERT	Ahn et al., 2004
PDGFRα	Pre-Chondral Mesenchyme	Cre	Roesch et al., 2008
		CreERT	Kang et al., 2010
		Flox	Tallquist et al., 2003
SOX9	Pre-Chondrocytes, Chondrocytes, and Enthesis	Cre	Akiyama et al., 2005
		CreERT	Kopp et al., 2011
		Flox	Akiyama et al., 2002
NKX3.2 (BAPX1)	Pre-Chondrocytes and Chondrocytes	Cre	Stanfel et al., 2006
HES1	Pre-Chondrocytes and Chondrocytes	Cre	Imayoshi et al., 2003
		CreERT	Kopinke et al., 2011
COL2A1	Chondrocytes, Stromal Cells	Cre	Ovchinnikov et al., 2000
		CreERT	Nakamura et al,. 2006
ACAN	Pre-Chondrocytes and Chondrocytes	CreERT	Henry et al., 2009
COL10A1	Hypertrophic Chondrocytes	Cre	Chen et al., 2019
RUNX2	Hypertrophic Chondrocytes	Cre	Rauch et al., 2010
		Flox	Tu et al., 2012
IHH	Hypertrophic Chondrocytes	Flox	Razzaque et al., 2005
PTHrP	Resting Chondrocytes	CreERT	Mizuhashi et al., 2018
		Flox	He et al., 2001
VEGF	Hypertrophic Chondrocytes	Flox	Gerber et al., 1999
MMP13	Hypertrophic Chondrocytes	Flox	Stickens et al., 2004
GDF5	Articular Cartilage (Interzone), Enthesis	Cre	Rountree et al., 2004
		CreERT	Shwartz et al., 2016
WNT9A	Articular Cartilage (Interzone)	Flox	Spater et al., 2006
TPPP3	Synoviocytes	CreERT	Harvey et al., 2019
PRG4	Synoviocytes	CreERT	Kozhemyakina et al., 2015
		Flox	Hill et al., 2015
OSX	Perichondrium, Periosteum, Stromal	Cre	Rodda et al., 2006
		CreERT	Maes et al., 2010
HES1	Perichondrium and Periosteum	Cre	Imayoshi et al., 2003
		CreERT	Kopinke et al., 2011
DLX5	Perichondrium and Periosteum	Cre	Ruest et al., 2003
		CreERT	Taniguchi et al., 2011
OCN (BGLAP)	Osteoblasts	Cre	Zhang et al., 2002
		CreERT	Yoshikawa et al., 2011
OPN (SPP1)	Osteoblasts	Cre	Español-Suñer et al., 2012
DMP1	Pre-Osteocytes and Osteocytes	Cre	Lu et al., 2007
		CreERT	Powell et al., 2011
SOST	Pre-Osteocytes and Osteocytes	Cre	Xiong et al., 2015
		CreERT	Maurel et al., 2019
TRAP (ACP5)	Osteoclasts	Cre	Chiu et al., 2004
CTSK	Osteoclasts	Cre	Winkeler et al., 2012
		CreERT	Sanchez-Fernandez et al., 2012
LYSM	Myloid (Pre-Osteoclast)	Cre	Clausen et al., 1999
		CreERT	Canli et al., 2017
RANK (TNFRSF11a)	Myloid (Pre-Osteoclast)	Cre	Maeda et al., 2012
LEPR	Marrow Stromal Cells	Cre	Leshan et al., 2006
		Flox	Coehn et al., 2001
ADIPOQ	Marrow Adipogenic Progenitors	Cre	Eguchi et al., 2011
		CreERT	Jeffrey et al., 2015
CLEC11A	Marrow Osteogenic Progenitors	Flox	Zhang et al., 2021

Table 2:

Mesodermal Developmental Structures, Stages, and Markers

Mesodermal Subunit	Lineage Markers	Developmental Structures	Adult Structures
Mesoderm (Gastrula)	TBXT/Brachyury
Chordamesoderm	COL2A1 SHHA IHHB Cytokeratin 8/18/19 Galectin-3	Notochord	Vertebral Disks Intervertebral Bodies Nucleus Pulposus
Paraxial Mesoderm	MSGN1 (Posterior Paraxial) TBX6 (Anterior Paraxial) PAX3 (Anterior Paraxial) UNCX (Posterior Somite) TBX18 (Anterior Somite)	Somites	Axial and Appendicular Dermis Axial and Appendicular Skeletal Muscle Axial Ligament/Tendon Axial Articulation (Cartilage/Synovium) Axial Bones
Intermediate Mesoderm	PAX2 LHX1		Kidney Gonads
Lateral Plate Mesoderm	FOXF1 PRRX1	Somatic Coelum	Thoracic and Abdominal Cavities
		Somatic Mesoderm	Body Wall Limb Bud Appendicular Ligament/Tendon Appendicular Articulation (Cartilage/Synovium) Appendicular Bones Parietal Peritoneum Parietal Pleura Parietal Pericardium
		Splanchnic Mesoderm	Visceral Peritoneum Visceral Pleura Viasceral Pericardium Visceral Linings

2.1. The Mesodermal Lineage

The trilaminar gastrula develops axially-to-laterally around the neural tube into several distinct mesodermal structures.⁶ Centermost of these is the axial mesoderm, also known as the chorda-mesoderm, which underlies the neural tube and gives rise to the notochord.⁷ A majority of cells and structures derived from the axial mesoderm are transitory during development with only some components of the vertebral disks, intervertebral bodies, and nucleus pulposus present into adulthood.⁷

The next three mesodermal structures arranged from medial to lateral are the paraxial mesoderm, the intermediate mesoderm, and the lateral plate mesoderm (Table 2).⁸ Of these, the paraxial and lateral plate mesoderm are most relevant to musculoskeletal development lending to the generation of the somites and somatopleure respectively (Table 3).⁸ The intermediate mesoderm is not a primary contributor to musculoskeletal development instead developing into the urogenital system. From this is can be determined that lineage tracing of the musculoskeletal system must focus on expression of markers distinct to the paraxial (mesogenin 1 (MSGN1), T-Box transcription factor 6 (TBX6), delta 1 (DLL1), and paired box 3 (PAX3)) and lateral plate (forkhead box protein F1 (FOXF1), paired related homeobox 1 (PRRX1), and GATA binding protein 4 (GATA4)) mesoderm (Table 1&2).⁸

Table 3:

Somite Developmental and Adult Structures

Somite Derivative	Developmental Structures	Adult Structures
Dermomyotome	Dermotome Myotome	Dermis Skeletal Muscle
Sclerotome (Primordial)	Syndetome Arthrotome Endotome Sclerotome (Proper)	Tendon Joints Endothelial Bones

2.2. Paraxial Mesoderm and Somite Formation

The paraxial mesoderm is the source of the axial musculoskeletal system and has a unique pattern of forty-two to forty-four paired cranial to caudal segments called somites. After paraxial mesoderm first coalesces and elongates it develops from posterior to anterior with the posterior being more immature and the anterior more committed. Spatial restriction of MSGN1 and TBX6 marks the posterior somite while PAX 3 marks the more anterior component (Table 2).⁸ Along the craniocaudal axis elongation and differentiation are driven by wingless-type integration site family (WNT), neurogenic locus notch homolog protein (NOTCH), and fibroblast growth factor (FGF) signaling.^9–11 Mediolateral development and identity is in turn driven by bone morphogenetic protein (BMP) signaling gradients. Consequently, the anterior zone of the paraxial mesoderm segments into the embryonic somites.^9–11 Post-somite segmentation mesenchyme homeobox 1 (MEOX1) may be utilized to track the paraxial lineage (Table 1). As somites mature, WNT6 signaling in the dorsal ectoderm activates transcription factor 15 (TCF15/Paraxis) which organizes mesenchymal-to-epithelial transition of the anterior somite.^66–67 This generates a trilaminar structure consisting of a mesenchymal core with both dorsal and ventral epithelial lamina. The epithelial lamina eventually forms a cleft which separates the developing somite.^12–13 Once separated, somites compartmentalize along a dorsoventral and mediolateral axis in a gradient of the dermomyotome (further dividing into dermotome and myotome), syndetome, and sclerotome (Table 3). Each of these components are further subdivided into anterior and posterior zones marked by T-Box transcription factor 18 (TBX18) and UNC homeobox (UNCX) respectively.^14–16 The dermatome will later develop into the dermis, connective tissues, and adipose of the neck and trunk. The myotome will develop into axial muscles with a further subdivision in the epimere (extensors) and the hypomere (anterior thoracic and abdominal walls).^17–18 Continuing in this manner the sclerotome will give rise to the syndetome (tendon/ligaments), the arthrotome (articular cartilage), the endotome (some endothelial tissues), and the proper sclerotome which is origin for axial bones.

2.3. Lateral Plate Mesoderm and the Limb Bud

Expression within the early lateral plate is highly heterogeneous with individual markers either a) not marking all lateral plate lineages (FOXF1, PRRX1, GATA4) or b) marking non-lateral plate lineages (TBXT). Grossly, the lateral plate mesoderm divides into a dorsal somatic mesoderm (somatopleure) and a ventral splanchnic mesoderm (splanchnopleure).^19–20 Between these sheets is the intraembryonic or somatic coelom which gives rise to the abdominal and thoracic cavities. Of these, the somatopleure contributes primarily to the musculoskeletal system. The somatopleure gives rise to body wall and is subdivided further into the anterior limb bud, the posterior limb bud, and the non-limb forming wall including the parietal peritoneum, pleura, and pericardium. The anterior and posterior limb buds may be marked by T-box transcription factor 5 and 4 (TBX5, TBX4) respectively (Table 4). As the nascent limb bud begins to develop, cells derived from the somitic paraxial mesoderm migrate including myogenic populations which will later form the appendicular musculature. Polarity and positional signaling is a complex and highly conserved system, however, a full accounting of the early limb development process is beyond the scope of this review. As the limb begins to form along the gradient axes established during the early limb bud musculoskeletal structures beginning as early mesenchymal condensates will form which will develop later into established structures. Higher level patterning of these structures appears dependent on homeobox (HOX) gene transcription factors. During development the proximodistal axis of the limb divides into three regions which from proximal to distal include the stylopod (HOX9/10), the zeugopod (HOX11), and the autopod (HOX 12/13) (Table 4).^21–22 Bones which form from the stylopod include the humerus and femur. Those forming the zeugopod include the radius, ulna, tibia, and fibula. The carpus/tarsus, metacarpus/metatarsus, and phalanges form the autopod (Table 4).²³

Table 4:

Limb Development, Markers, and Adult Structures

Limb Development	Lineage Markers	Adult Structures
Forelimb Bud Mesenchyme	TBX5	Forelimb
Hindlimb Bud Mesenchyme	TBX4 PITX1	Hindlimb
Stylopod	HOX9/10	Humerus Femur
Zeugopod	HOX11	Radius/Ulna Tibia/Fibula
Autopod	HOX12/13	Carpus/Metacarpus Tarsus/Metatarsus Phalanges

3.0. Musculoskeletal Development:

3.1. Cartilage and Bone:

Initial formation of appendicular bone occurs by the process of endochondral ossification. A condensation of lateral plate mesoderm-derived mesenchyme, typically marked by platelet-derived growth factor receptor alpha (PDGFRα), forms in the area which has been patterned by HOX genes to form bone primordia (Table 1&5).^23–27 As the mesenchymal condensate expands cells which are more centrally located will transdifferentiate in chondrocytes. In a majority of pre-chondrocytes this event is marked expression of the master chondral regulatory SRY-box transcription factor 9 (SOX9) followed by the osteogenic inhibitor Nirenberg and Kim 3 homeobox 2 (NKX3.2).^34–35 Peripheral to this, a separate SOX9-negative group of pre-chondrocytes expressing hairy and enhancer of split-1 (HES1) begins to emerge. At this time the surrounding mesenchyme will also begin to develop into the multilaminar perichondrium (Table 5).^28–29 Closest to the chondral template, in the deepest perichondral layers multi-potent osterix (OSX) positive cells can be found with both potential as both chondrogenic and osteogenic precursors.^30–31 A further HES1 and distal-less homeobox 5 (DLX5) positive population stratifies to the superficial perichondrium which will later contribute both osteoblasts and later marrow stem cells.^32–33 The pre-chondrocytes then begin to mature and express chondroid matrix markers including collagen type 2 alpha 1 chain (COL2A1) and aggrecan (ACAN). In the deepest portion of the cartilage template chondrocytes will begin to mature then hypertrophy.^36–37 These hypertrophic chondrocytes begin to express collagen type X alpha 1 chain (COL10A1) as well as vascular signals such as vascular endothelial growth factor (VEGF) to trigger surround vasculature to invade the chondral template (Table 5).³⁸ Vascular invasion triggers both simultaneous resorption of cartilage and ossification as well as the generation of the marrow space.

Table 5:

Appendicular Musculoskeletal Structures and Derivatives

Appendicular Musculoskeletal Component	Lineage Markers	Mesodermal Lineage
Pre-Osseous Mesenchymal Condensate	PRRX1 PDGFRa	Lateral Plate
Chondral Template/Immature Chondrocytes	SOX9 NKX3.2 HES1 COL2A1 ACAN	Lateral Plate
Hypertrophic Chondrocytes	COL10A1 RUNX2 VEGF NOTCH 1/2 IHH MMP13	Lateral Plate
Perichrondrium	OSX HES1 DLX5 WNT5A WNT11	Lateral Plate
Trabecular Bone (Osteoblasts)	OCN OPN COL1A1	Lateral Plate
Osteoclasts	TRAP Cathepsin K	Lateral Plate
Articular Cartilage	GDF5 GDF6 WNT9A TPPP3	Lateral Plate
Synovium	PRG4	Lateral Plate
Tendon/Ligament	SCX MKX	Lateral Plate
Enthesis	SCX SOX9 GDF5	Lateral Plate
Myotendinous Junction	COLXXII	Mixed
Muscle	PAX3 MYF5	Paraxial
Muscle Satellite Cells	PAX7 NOTCH	Paraxial

As bone matures a highly diverse mesenchymal and hematopoietic niche develops in the medullary space such that common lineage markers may represent a myriad of distinct sub-populations. Hematopoietic stem cells (HSCs) are incredibly heterogenous such that a combination of lineage positive and negative markers are required to identify specific populations.^39–40 One common combination includes signaling lymphocytic activation molecule 1 (SLAMF1/CD150) and stem cell antigen-1 (SCA1) positivity alongside SLAMF2/CD48 and lineage (LIN: CD3, CD45R (B220; PTPRC), CD11B, Ly6G, TER119) negativity.^39–40 Some cre-driven lineages have also been developed which included FMS like tyrosine kinase 3 (FLT3), HOXB5, and alpha-catulin (CTNNAL1).^39–40 In addition to hematopoietic stem cells marrow is enriched for leukocyte populations (monocytes/macrophages, dendritic cells), megakaryocytes, endothelial, and neuronal tissues as well as resident mesenchymal stromal cells (MSCs).^39–40 During early post-natal development nestin (NES) marks both marrow stromal cells and osteoblasts. These marrow stromal cells include several mesenchymal progenitor populations (OSX, COL2A1, HOXA11) which modulate osseous homeostasis.^39–40 Within the early metaphysis OSX and HOXA11 positive population generate a stable leptin receptor (LEPR) expressing stromal population which includes the presumptive skeletal stem cells.^41–42 These LEPR positive populations may then subdivide into an adiponectin (ADIPOQ) positive and osteolectin (CLEC11A) positive lineage of adipocytes and osteogenic progenitors respectively (Table 1&5).^43–44 Recent data suggests that the spatial relation of these LEPR positive populations (perisinusoidal vs. periarteriolar) may help mark those cells of adipogenic (perisinusoidal) vs. osteogenic (periarteriolar).⁴⁴

3.2. Growth Plates:

Upon formation of the initial or primary ossification center the immature chondrocytes which remain proximally and distally become aligned and organize into zones of rest, proliferation, and hypertrophy. These parallel disks and the primary ossification center when taken as a unit are referred to as the epiphyseal or growth plate. This is an ultimately transient structure that is maintained by negative feedback of parathyroid hormone-related protein (PTHrP) secreted by resting chondrocytes which in turn binds reciprocal receptors on proliferating chrondrocytes in a paracrine manner.^45–47 Once bound this suppresses terminal differentiation to the indian hedgehog (IHH) producing hypertrophic chondrocytes. As this is a paracrine effect, sustained proliferation of chrondrocytes eventually moves away from the resting zone, reducing the PTHrP signal and IHH signaling increases resulting in activation of runt-related transcription factor 2 (RUNX2) supporting terminal osteogenesis (Table 1&5).

3.3. Joints/Articulations:

Articular surfaces form at the surface between the bones of developing limb fields and the earliest histologic precursor to joint formation is the development of an interzone between distinct cartilage templates. These are avascular zones of densely packed mesenchymal cells which are flattened and oriented perpendicular the adjacent chondral templates. Several markers have been identified which may mark interzone cells joint formation. WNT9A and growth differentiation factor 5/6 (GDF5/6) are some of the earliest consistent markers of interzone cells consequently defining an articular cartilage lineage.^48–51 Interzone markers appear to have complex and overlapping expression patterns driven by both positive (odd-skipped-related transcription factor 1/2 (OSR1/2) and NOG) and negative regulatory elements (BMP 2/4 and fibronectin receptor/alpha5beta1 integrin (FNR)).^51–55 Synovial demarcation can further be identified by expression of tubulin polymerization-promoting protein family member 3 (TPPP3).⁵⁶ TPPP3 expression aligns with cavitation and alongside proteoglycan 4 (PRG4/lubricin) mark synoviocytes (Table 1&5).⁵⁶

3.4. Ligaments, Tendons, and their Interfaces:

Tendons and ligaments are force-transmitting structures which interface between muscle and bone or bone and bone respectively. The developmental differences between tendons and ligaments are poorly defined and for the sake of this review they will mostly be discussed as a single developmental entity. Tenocytes and ligamentocytes are the primary cell of the tendon and ligament respectively and are of mesodermal lineage. Tenocytes of the appendicular and axial body have different origins with those of the trunk arising the syndetome, a somite sub-element derived from the paraxial mesoderm.⁵⁷ In the limb, tenocytes are derived from the mesenchyme of the lateral plate mesoderm and differentiate in response axial-to-appendicular migration of myoblasts from the myotomes. Currently it is believed that tenocytes of either origin are dependent on expression of scleraxis (SCX) and mohawk (MKX) during differentiation (Table 1&5).^58–62 Consequently, a majority tenocytes are presumed to be of a SCX-lineage regardless of paraxial or lateral plate source. This is not universal, however, and some anchoring ligaments as well as short tendons are not entirely dependent on SCX expression. Once tenocytes have developed tendon development occurs in two stages. During the first stage, ribbons of tenocytes delineate a zone of collagen fibril deposition. This is followed by a second mohawk homeobox (MKX) dependent second stage in which the collagen template is utilized to grow and develop mature tendons. In force-transmitting tendons and ligaments this process is presumed to be related to mechanical stimulation.

Where tendons and ligaments connect to bones, we find a distinct transitional structure termed the enthesis. This is a zonal structure primarily composed of fibrocartilage which gradually transitions between the primarily fibrous elements of the tendon/ligament to the primarily mineralized elements which transition to bone. The fiber-cartilage-bone transition of the enthesis shares some histologic similarities with endochondral ossification at the cartilage-bone interface. Interestingly, a gradient of overlapping expression including SCX, SOX9, GDF5, and hedgehog are critical for both identity and maturation.^63–65 The exact signaling process is as of yet unknown, however as with ligaments and tendons proper maturation of the system is presumed to be dependent on transmission of mechanical signals.

Opposite to the enthesis is the myotendinous junction. This is the interface between muscle and tendon or more accurately between the investing fascia of muscle and the tendon-proper. As opposed to the gradual transition of the enthesis the myotendinous junction resembles long extended processes each sheet of extracellular matrix between sarcolemma units of muscle directly conjoins together and into tendon collagen fibers.⁶⁶ While maturation of this structure is well described in adults the exact developmental origin of the myotendinous junction is unclear. Collagen XXII is one of the few proteins primarily restricted to the myotendinous junction and thus may be helpful in marking fate of resident cells.⁶⁷

3.5. Muscles:

Skeletal muscle, whether axial or appendicular, is derived from the paraxial mesoderm. Skeletal muscle first arises with the determination of pre-myogenic progenitor cells and eventually myoblasts in the somite. Earliest myogenesis initiates with activation of first PAX3 then myogenic factor 5 (MYF5) in the dorsomedial somite.^68–69 Embryonic or primary myogenesis involves the production myofibers from PAX3 positive dermatomyotome progenitors.^70–71 Myf5 then marks the primary myotome present immediately ventral to the dermomyotome proper. Mononuclear myocytes arise in this zone gradually elongating along a craniocaudal axis to span the somite. Simultaneously cells from the dermomyotome continue to translocate to the myotome acting as myogenic progenitors while more laterally myogenic cells migrate to the lateral plate and into limb bud to form appendicular skeletal muscle. These appendicular myocytes will then coalesce and form myofibers of their own. This process of primary myogenesis forms the primordial template for each of the skeletal muscles of the trunk and limbs.^72–73 Secondary myogenesis occurs as further PAX3 positive progenitors express PAX7 then translocate to fuse to the primary fibers.⁷⁴ Growth throughout this stage involves cell fusion rather than hypertrophy as is described in adult tissues. A portion of the PAX7 positive progenitors from the central dermatomyotome will maintain their developmental characteristics under Notch signaling and will contribute to the satellite cell population of adult muscle stem cells.

The transition from mononuclear myoblast to myotube to myofiber begins with recognition and adhesion of adjacent myoblasts. Elongation is formed by the successive fusion of chains of myoblasts and or multiple myotubes. Myotubes then undergo maturation in a process call myofibrillogenesis in which the myotube begins to segment into contractile modules called sarcomeres which are surrounded by and anchored to specialized connective tissue called the endomysium. In addition to mechanically supporting the sarcomere the endomysium transmits neuronal signals. This segmentation gives rise to the myofibrils present in adult muscle which extend the length of the muscle and are anchored by the myotendinous junctions.

4.0. The Adult Musculoskeletal System

Here we will discuss the process of musculoskeletal healing with a focus on those tissue- and site-specific markers which may be utilized as reporters during or after the healing process.

4.1. Fracture Healing, Bone Homeostasis, and Pathologic Bone Formation:

Adult bone is inhabited by several cells directly involved in osseous homeostasis including osteocytes/osteoblasts (osteocalcin (OCN), dentin-matrix protein (DMP1), sclerostin (SOST)) and osteoclasts (tartrate resistance acid phosphatase (TRAP)) as well as those of the periosteal and endosteal linings and the medullary/stromal space (Leptin Receptor (LEPR), neuron-glial antigen 2 (NG2)/chondroitin sulfate proteoglycan 4 (CSPG4), NES, Gremlin 1 (GREM1)).^39–44 Extent of injury helps define which cellular populations are involved in repair. While microfractures can be managed during normal bone homeostasis significant fracture sets off an organized cascade of steps which utilizes the cellular and biochemical machinery of endochondral ossification to restore osseous integrity. As endochondral ossification utilizes developmental machinery, mesenchymal (PDGFRα), early chondral (SOX9, COL2A1), and hypertrophic cartilage (COL10A1, RUNX2) markers are valuable to track new versus pre-existing bone. Bone healing typically begins at time of fracture with immediate hematoma formation secondary to rupture of periosteal and endosteal vessels as well as disruption of the marrow cavity (Fig. 1A). The subsequent clotted hematoma serves as both scaffold and signaling center for leukocytes, endothelial cells, and ultimately mesodermal lineage cells. Disruption of the medullary canal and consequent sinusoidal space allows for activation of megakaryocytes, leukocyte populations, endothelial cells (platelet endothelial cell adhesion molecule 1 (PECAM1/CD31) and apelin (APLN)), and mesenchymal stromal cells (LEPR (Fig. 1B&C), stem cell factor (SCF), and C-X-C motif chemokine ligand 12 (CXCL12)) responsible of skeletal and hematopoietic maintenance.^39–44 Ossification then occurs by two simultaneous processes. First, endochondral ossification will proceed from within the callus as described above. Second, at the level of the periosteum woven bone will be formed directly from resident osteoprogenitors. OSX-lineage positive cells in the periosteum are noted to proliferate and co-localize to/co-invade alongside vascular populations contributing to woven bone (Fig. 1B&C).^30–31 Over time immature bone will remodel until the initial injury is functionally resolved.

Figure 1:

(A) Stages of fracture repair. (B) inducible cre strategy using an osx-cre-ERT2 transgenic mouse crossed with a Rosa-tdTomato mouse. After tamoxifen injection, osterix-expressing cells will express tdTomato, as shown here in post-operative day 3 section of a tibial injury. Immunofluorescence against osterix revealed co-labeling of osx and tomato-positive cells. (C) Non-inducible labeling strategy utilizing a LepR-cre and tdTomato mouse in which all LepR-expressing cells and their progeny will express tdTomato.

The body retains capacity for de novo bone and cartilage formation well into adulthood. While this can be physiologic there are several conditions of pathologic bone formation. One common variant is the formation of osteophytes. Inflammatory injury to periosteum or synovial lining stimulates proliferation of mesenchymal stem-like cells present in the periosteum stimulating rapid proliferation.^75–76 These cell masses may then undergo endochondral ossification and ankylosis to the original bone.^75–76 During this process intramembranous ossification may occur and contribute to the final osteophyte. A less common but potentially debilitating process is heterotopic ossification or ectopic bone formation.⁷⁷ Similar to fracture healing, this condition can be traced by the same genes which mark each stage of endochondral ossification in which a mesenchymal anlage undergoes chondrification followed by ossification and eventually maturation as described above.^77–78

4.2. Articular Cartilage Injury and Repair:

Articular cartilage is a well-organized, stratified/zonal structure. Articular cartilage is primarily avascular and receives nutrient exchange superficially by synovial fluid and at the deep surface by subchondral bone. Developmental lineage markers including SOX9 may be retained into adulthood, however, joints may also be effectively marked by ubiquitous chondrocyte extracellular matrix markers (COL2A1, ACAN, Cartilage Matrix Protein (Matrilin)).⁷⁹ In mature articular cartilage expression patterns of many extracellular matrix components are zonally restricted. Aggrecan is primarily limited to those less-mineralized chondrocytes towards the articular surface away from the ‘tidemark’.⁸⁰ Matrilin-3 localizes to joints and is further restricted to tangential and upper middle cartilage zone chondrocytes.⁸¹ Proteoglycan 4 (PRG4)/Lubricin is expressed more abundantly in chondrocytes of the superficial zone as well as synoviocytes.^82–83 Other cartilage matrix proteins are less specific with Matrilin-1 present primarily during development and more in adult joints whereas Matrilin-2, while common in adult joints is also present in several extraskeletal tissues.^84–85

Cartilage healing is limited under physiologic conditions. Given the relative avascularity of the superficial zones, partial-thickness injuries often demonstrate poor healing. Full-thickness defects, however, with disruption of subchondral bone are more able to demonstrate a limited form of regeneration. Under these conditions a fibrocartilaginous anlage similar to that seen in endochondral ossification forms. Healing at this level involves recruitment of GDF5 (articular) and NES (skeletal) lineage progenitors to form an osseocartilaginous callus, however, in absence of additional trauma or inflammation this typically arrests without further ossification.^85–87 Repetitive injury/healing cycles such as seen in osteoarthritis (OA) further degrade the articular surface and can tip the balance towards osteogenesis as described below.

4.3. Tendinous and Ligamentous Healing:

Tendons and ligaments share similar pathways to healing, each undergoing overlapping stages of hematoma, inflammation, proliferation, and remodeling after trauma. Endothelial injury, relative hypoxia, and breakdown of coagulation byproducts signals for both endothelial and immunologic infiltration. Stages of inflammation include rapid response neutrophils followed by macrophages and then fibroblasts. Macrophages as well as resident cells of the epitenon and endotenon are key mediators of fibroproliferative signaling, expressing high levels of connective tissue growth factor (CTGF), transforming growth factor beta (TGFβ), and insulin-like growth factor 1 (IGF1) throughout the transition from inflammation to proliferation.^88–89 The epitenon is a critical source for tendon progenitor cells (PDGFRA, Tubulin polymerization-promoting protein family member 3 (TPPP3), NES, SCX) with both perivascular and resident fibroproliferative cells contributing to tenogenic healing.⁹⁰ At the enthesis additional contributions from tendon, bone, and the fibrocartilaginous anlage contribute to repair of the bone-cartilage-fibrous transition. In neonates, glioma-associated oncogene/zinc-finger protein (GLI1) marks a resident progenitor population capable of replenishing mineralize fibrocartilage (Table 1).⁹¹ This population loses Gli1 expression in adulthood and terminally differentiates with reduced regenerative capacity after injury.

4.4. Muscle Homeostasis, Healing, and Degeneration:

Muscle healing is highly dependent on the degree of injury and similar to bone is a balance of ordered regenerative processes and more fibrotic terminal processes in the cases of extensive or critically sized injuries. Repair after injury is typically managed in stages of degeneration followed by regeneration. Degeneration occurs after any injury sufficient to disrupt myofiber and consequently sarcolemma integrity. This results in rapid necrosis of affected myofibers and both signals for and is followed by inflammatory infiltration by phagocytic populations – first neutrophils then predominantly macrophages. Muscle regeneration follows and in healthy tissues this is typically marked by myogenic proliferation in which new myofibers are formed from the division, differentiation, and fusion of fibers to injured tissue. Injury additionally induces M-Cadherin expression in both injured myotubes and in satellite cells and M-Calpain in fusing myoblasts.⁹² New myofibers will then undergo hypertrophy prior to transition back to a morphologically consistent state with surrounding muscle.

Satellite cells are undifferentiation myogenic cells located peripheral to mature myofibers. These are mononuclear cells typically arrested under NOTCH signaling in a progenitor state and are most commonly derived from PAX7 positive populations from the central dermatomyotome.^93–94 When muscle is uninjured and at rest these cells are typically quiescent and are only activated in response to trauma or high demand. Satellite cells are marked by myocyte nuclear factor (MNF) signaling.⁹⁵ MNF is distinct in that it has two distinct MNFα and MNFβ with the beta isoform more abundant when quiescent and the alpha isoform more abundant during satellite cell activation. Satellite cells are capable of stem-like behavior and when activated the pool will asymmetrically proliferate with a portion their progeny differentiating to generate new myoblasts with the capacity to fuse during regeneration. Asymmetry of this process is thought to be necessary for maintenance of the stem cell population and is related to the variability in ratios between NUMB and NOTCH1 signaling in daughter cells.^93–96 Upon activation satellite cells will express MYF5, myogenic differentiation 1 (MYOD), and signal transducer and activator of transcription (STAT3) during proliferation followed by myogenin 1 (MYOG1), myogenic factor 6/myogenic regulatory factor 4 (MYF6/MRF4), and Snail family transcriptional repressor 2 (SNAI2) during differentiation and fusion.^97–100 Given the range of factors associated with muscle regeneration several animal models now exist with which to aid in the study and tracking myocyte and satellite cell populations (Table 1). In addition to satellite cell activity, current literature supports contribution from circulating NES-positive MSC and HSC progenitors.^101–102

When myogenic regeneration fails, volumetric muscle loss and/or replacement occur. Muscle fibrosis, fibrofatty change, and sarcopenia make up a cluster of diseases related in part to the failure of normal muscle homeostasis. In both muscle fibrosis and fibrofatty change invasion and proliferation of non-myogenic cells are key to pathology with mesenchymal tissues (PDGFRA, SCA1).^103–105 Because of the similarities between the two processes a significant body of work has gone into the identification of a common fibro/adipogenic progenitor. Extent and type of injury guides fibrotic vs. adipogenic (peroxisome proliferator- activated receptor gamma (PPARγ) and glycogen synthase kinase 3 (GSK3)) differentiation.^106–108 Intra-muscular ectopic ossification, as described in heterotopic ossification above, may incorporate this fibro/adipogenic progenitor population, however, tracing experiments suggest alternate cell sources for these lesions.^109–110

5.0. Technical Aspects of Lineage Tracing and Fate Mapping:

The musculoskeletal system is marked by its diversity of adult and developmental structures. Because of the transient nature of processes such as somite differentiation, endochondral ossification, or satellite cell activation, lineage tracing of population across time is necessary to study the system as a whole.^111–112 Consequently, tracking a cell or cellular lineage across time requires the ability to mark the target in such a way that we are able to accurately identify either the original cell or its progeny at a later date. Though this may be achieved by direct observation in physically small, developmentally immature, or translucent specimens, markers become critical in the study of health and disease in larger, adult, or mammalian models.^114–116 Accuracy of any such technique is dependent on several key factors. First, the marker must persist to the desired generation without loss of sufficient signal. Second, the marker may not be spread or in some manner diffuse to cells not of the desired lineage. Thirdly, the marker must not affect the normal proliferation, differentiation, migration or any other part of the cell lines development in such a way as to alter the normal function of the cell or its progeny. This can be achieved in a variety of ways which we will briefly summarize below.

5.1. Model Organisms

Historically invertebrates and lower vertebrates have been utilized for fate mapping and lineage tracing as their small size, availability, and rapid generational turnover has facilitated such endeavors. In the modern era, however, more and more clinically translatable studies draw data from mammalian models. Rodents, as example, are logistically amenable to both germline mutations and study of both developmental and pathologic processes. A wide range of existing reporter animals now exist for commercial purchase or generation with high degrees of accuracy (Table 1).^117–118 In this review we will not focus on individual strains of animals, but rather those genes which, during development or disease, are relevant to lineage tracing of musculoskeletal cells. Briefly, it is important to note that despite the highest degree of clinical relevance, direct genetic labelling in humans is not supported in the setting of current legislative and bioethical concerns relating to germline modification. More available are those nuclear and mitochondrial somatic mutations which occur naturally and allow for post-differentiation characterization of shared lineages between cells and tissues, or in the labelling of tissues in vitro for autotransplantation. Given the relative infancy of this field, however, no comprehensive lineage analysis is present for the musculoskeletal system by these techniques.

5.2. Direct and Indirect Labeling

The addition of a dye to a cell population is one of the earliest techniques utilized for temporo-spatial cell tracking.^{114,119–120} Non-toxic or vital dyes were critical to the earliest embryonic fate mapping, and successive improvements in dye techniques have led to a broad range of lipid- and water-soluble and insoluble dyes as well as radiolabeled or fluorophore-conjugated compounds capable of sustaining signal across several generations of cell division. Dyes can be taken up and trapped in cells, incorporated into DNA, or can irreversibly bind to lipids and/or proteins of the cell membrane and several commercially available products exist to make use of these various mechanisms. Depending on the technique, dyes can additionally be used for pulse-chase experiments to track cell division and or to identify cell populations in which no genetic marker is currently known.^121–122 Even modern dyes have some limitations. Because dyes are limited to their original exposure signal strength can be lost with each cellular division. Consequently, cells which divide rapidly are more likely to deplete their signal. Furthermore, particularly with those dyes that bind to proteins or lipids there is the risk that either the dyed cell or other phagocytic cells will recognize the foreign material which may lead to signal degradation, changes to cell function, or even cell death.

5.3. Gene Constructs

Modern lineage tracing primarily utilizes one or more genetic constructs to generate a stable and selective reporter. The actual insertion of these constructs into organisms whether germline or in adult cells and whether by mechanical^123–125, electrical¹²⁶, chemical¹²⁷, viral (adenoviral, lentiviral)¹²⁸, or proteomic (Zinc Finger Nucleases, TALENs, CRISPR/Cas9)^129–131 is beyond the scope of this review in which we will focus primarily on specific constructs and their application to lineage tracing in the setting of musculoskeletal development.

Historically lineage tracing utilizing gene constructs is dependent on a system of inducible insertion/deletion centered around Cre-Lox recombination described by Sternberg et al. Successive iterations have let to the identification of a range of site-specific recombinases and more refined control of recombination events.^132–133 In brief, a promoter, typically that of a specific gene of interest which expression limited to an ancestral cell or cells within the lineage of interest, is linked to the production of a site-specific recombinase. In cells which produce this recombinase a second, typically inert, transgene flanked by some identifying sequence undergoes a recombination event to activate said transgene and thus marks the cell (Fig. 2A, B).¹³⁴ These transgenes are often colorimetric, genetic, functional, or fluorescent reporters, described later in this review, which will remain present and active into the progeny of a given cell consequently marking all descendent in a specific lineage. As these reporter cassettes and their activating promoters become more complex, systems of alternating color can be used to mark extensive and cellular lineage with ever increasing accuracy (Table 1).

Figure 2:

Schematic depicting the difference between a (A) global knockout and a (B) conditional knockout strategy. (C) Inducible conditional knockout using the cre-ERT2 system, in which the cre recombinase is translocated into the nucleus only after administration of tamoxifen. (D) Cell ablation using an inducible conditional strategy in which the toxin receptor is expressed in cells after cre recombination excised the stop codon located adjacent to the coding sequence of the receptor. Once the toxin is delivered systemically, only cells that express the toxin receptor will be ablated. (E, F) Tetracycline-responsive elements either activate or inactivate the cre recombinase, resulting in inhibition or activation of downstream gene expression.

The Cre (P1 bacteriophage cyclization recombinase) of the Cre-LoxP system is one of a family of tyrosine recombinases which includes other commonly utilized Flp (flippase),¹³⁵ Dre (D6 site-specific DNA recombinase),^136–137 and Vika recombinases (Fig. 3A, B).^138–139 A tyrosine recombinase functions such that the enzyme targets a specific sequence of base pairs such as LoxP (Cre), Frt (Flp), rox (Dre), or vox (Vika). Base pair sequences targeted by tyrosine recombinases have a directionality determined by palindromic repeats centered around an asymmetric core. Specific recombinases then utilize these targeting sequences to guide recombination events. Cre-LoxP systems target paired LoxP sites with the intervening genomic locus being described as floxed. Location and orientation of LoxP sites can lead to excision (unidirectional, cis-placement of LoxP flanking sites), inversion (bidirectional, cis-placement of LoxP flanking sites), or translocation (trans placement of LoxP sites) with other tyrosine recombinases functioning similarly.¹⁴⁰ Because the enzymes in these systems target specific and typically non-overlapping target sequences, complex arrays of specific recombination events can be utilized to alter transcription of one or more genes of interest with increasing specificity. Variations of this technology have been described such as in the setting of DeaLT (dual-recombinase-activated lineage tracing) described by He et al., 2017 in which the Dre-Rox system is utilized to improve specificity of Cre-LoxP system for more precise fate mapping.¹⁴¹ Tyrosine recombinases are not the only option for site-specific recombination, though benefit significantly from not needing accessory factors for activation. Integrases, both tyrosine and serine based have been described with the PhiC31 serine integrase having been described for use in mammals.¹⁴² As opposed to the multidirectional tyrosine recombinases, serine integrases are unidirectional and function only at a limited range of genomic acceptor sites. This is beneficial during initial insertion of transgenes as there are more limited risks of off-target insertion.

Figure 3:

(A) Schematic of different recombinase strategies that result in either (B) deletion, (C) inversion or (D) translocation.

5.4. Inducibility, Ablation, and Transplantation

A benefit and limitation of traditional promoter-linked recombinases is that the initial stimulation of the promoter generates a cascade leading the transcription and activation of the recombinase and then essentially permanent change post-recombination. This is effective when the goal is to mark a lineage from first presence of a specific protein, however, limits specificity across the axis of time and in cases where mutation may induce embryonic lethality can lead to failure of the model as a whole. Numerous techniques have been developed to account for this by introducing a requirement for a second-agent at various stages of the recombination process (i.e., generation of a conditional transgene). One powerful example is incorporation of tetracycline or doxycycline-sensitive elements of the bacterial tet-operon which undergo conformational changes to either turn on or off (Tet-On, Tet-Off) in the presence of tetracycline/doxycycline (Fig. 2E, F).^143–144 Mutant recombinases provide further control as is the case of the inducible estrogen receptor (ER) transgene (Cre-ERT2) which allows for activation by the synthetic estrogen receptor ligand tamoxifen (4-hydroxytamoxifen) but not by physiologic 17β-estradiol (Fig. 2C; Table 1).^145–146 In each of these cases the absence of the conditional agent at baseline allows for exogenous control of the system – albeit at the cost of any unwanted side-effects associated with drug administration.

Depending on the promoter and/or recombinase/integrase, a single cell, an anatomic region, or a whole organism may be marked. This allows for a wide range of techniques for the modelling of development, health, and disease. Traditional techniques track the proliferation and differentiation of tissues within a single organism which, through elegant use of modifiable systems such as Cre-Lox and Flp-Frt activation and deactivation of markers in time may be utilized to track specific developmental or pathologic events. More complex constructs may be incorporated such as toxin receptors for cell ablation (diphtheria toxin, herpes simplex virus 1 thymidine kinase (HSV-1-tk)) or other functional proteins and or receptors which alter the actions and biology of the targeted cells (Fig. 2D).^147–150 In organisms where germline modification is not easily amenable alternate techniques such as cell or tissue transplantation or parabiosis may provide additional information.¹⁵¹ Parabiosis, as example, allows for the direct connection between the bloodstream of two separate organisms such that one with a reporter may supply circulating cells to another without the recipient carrying any mutation. This allows for further discrimination between local, regional, and circulating cell population.^152–153

5.5. Protein Reporters

Currently most forms of lineage tracing or fate mapping are dependent on the use of some form of label or reporter. A prerequisite for successful reporter selection is a predictable pattern of expression and/or retention and the absence of off-target effects that may unduly influence the surrounding system. The introduction of new peptide or protein can most simply be assessed with traditional histologic or immunoblot techniques, however, it is common with protein reporters to provide some visual cue for ease of assay without potentially destructive second-stage procedures. Historically 2-staged enzymatic reactions were used to generate a colorimetric effect, as seen in LacZ (β-Galactosidase) or GusA (β-Glucuronidase) with a variety of substrates providing different final colors.¹⁵⁴ Luminescent effects can be achieved with Firefly or Renilla luciferase in presence of luciferin or coelenterazine respectively which can then be utilized in a stoichiometric manner to determine activity in conditions of in vivo imaging.¹⁵⁵ Generation of mouse lines with pre-existing gene-trap loci such as the Gt(ROSA)26Sor strain developed in the 1990s further simplified the insertion of reporters allowing for the expansion of targeted animal models.^156–157 These advances set the stage for the use of endogenous fluorescent proteins. Fluorescent reporter constructs improve of the spectrophotometric specificity of luciferase-based constructs and allow for simultaneous live assay across a range of fluorophores. Early fluorescent proteins were modified from xenologous sources such as green fluorescent protein (GFP: Aequorea jellyfish) and red fluorescent protein (RFP/dsRed: Discosoma coral).^158–159 These have since been further modified directly or in the setting of fluorescence resonance energy transfer (FRET) between paired fluorophore constructs to generate a rainbow of fluorescent options which may be used to better discriminate tissues.^160–162 Further improvements to these techniques include dual and multi-color transgenic labeling systems.^163–165 One of the earliest of these was the mTmG developed by Muzumdar et al., in 2007 which expresses membrane-targeted Tomato (mT) in absence of Cre and transitions to membrane-targeted green fluorescent protein (mG) post-Cre recombination.¹⁶⁶ Multi-copy single-cassette multi-color fluorescent reporters were introduced with the Brainbow technology developed by the Lichtman group in 2007.^167–168 Initially based on the use of serially-fluorescent proteins arranged in tandem after a single promoter with expression controlled by cre-mediated deletion, successive iterations of this technology have accounted for cre-mediated inversion and a greater array of colors. Multiple copies of an injected transgene can activate expression of a wide range of fluorophores. Later iterations of this include the now well described Confetti Mouse - which utilized nuclear GFP, cytoplasmic yellow fluorescent protein (YFP) or RFP, and membrane bound cyan fluorescent protein (CFP) to provide a range of expression options within a single cell.¹⁶⁹ This Confetti mouse provides a similar applicability to the earlier mTmG animals in that a traditional Cre-reporter system can be crossed without requiring extensive back-crossing or application of further transgenic modification.

5.6. DNA Reporters

Lineage tracing may additionally be performed in absence of the production of protein markers. Variant form of lineage tracing utilizing the tracking of cells by artificial or inherited DNA sequences.¹⁷⁰ Artificial gene-constructs called ‘barcodes; may be inserted via the same techniques used to generate promoter/reporter constructs described above. One example of this is Polylox described in Pei et al., 2017 which utilizes the Cre-lox system for insertion of ten loxP sites spaced with alternating orientations.¹⁷¹ This technique allows for extensive excision/inversion events and consequently barcode diversity sufficient to allow for single cell fate mapping.

6.0. Conclusions

Lineage tracing in health and disease is a highly complex process which combines developmental biology, a clinical understanding of anatomy, physiology, and pathology with novel tools for cell tracking. Understanding the developmental processes that give rise to adult structures provides a guide for examining both regenerative and pathologic changes to adult tissue. The musculoskeletal system is composed of a wide range of highly specialized and functionally divergent tissues stemming from common mesodermal ancestors. Decades of research have generated a library of hundreds of genetic markers, some specific and some overlapping which mark the fate of cells belonging to each specific developmental lineage. Currently, lineage tracing provides a powerful tool with which to track these cells, however, the process is slow and limited by the availability of constructs and the heterogeneity of cell populations at each stage of development. Currently, the gold standard in lineage tracing remains models such as CreERT, Tet, and flippase systems, however, expansion of available animal models, development of novel methods for construct insertion such as CRISPR, and new systems for lineage tracing such as Barcode and Polylox continue to expand our ability to ask and answer questions with scientific and clinical relevance to better improve our understanding of musculoskeletal health and disease.