Electron Microscopy in the Study of Myofibroblastic Lesions

Brian Eyden, PhD

• Electron microscopy in the diagnosis and academic study of myofibroblastic lesions is discussed. from and tumor stroma are regarded as the nearest equivalent to a "normal" population with which to define myofibroblastic differentiation in tumoral and pseudotumoral lesions. Histological features include a plump-spindle-cell morphology, with an ill-defined cytoplasm paler and less fibrillar than in smooth-muscle cells, and matrix . Myofibroblasts stain for a-smooth-muscle , fibronectin, and . is found in some lesional myofibroblasts. The main ultrastructural features are prominent rough endoplasmic reticulum, modestly devel­ oped with focal densities ("stress fibers''), and fibronexus junctions. The latter are foci on the cell surface where intracellular myofilaments and extracellular fibronectin filaments converge. Myofibroblastic lesions vary in the extent to which they mirror this overall phenotype. Hypertrophic , Dupuytren's disease, nodular fasciitis, the fibroma­ toses, and inflammatory myofibroblastic tumors have the most developed myofibroblastlc features. , postoperative spindle-cell nodule, and fibroma of sheath are less well differentiated. Myofibroblastoma is among many lesions described as myofibroblastic which, however, appear to show a kind of smooth-muscle differentiation. Some spindle-cell malignancies express myofibroblastic features. Copyright2003, Elsevier Inc. All rights reserved.

INDEX WORDS: Myofibroblast, pathology, electronmicroscopy, fibronexus, fibronectin-fibril

OMPAREDWITH MOST CELLS of the connec­ for example, SlOO protein and melanosomes in malig­ C tive tissues (the , smooth-, nant melanoma, and a-sarcomeric actin and , and so on) for which tumoral equivalents have in rhabdomyosarcoma. For myofibroblastic lesions, been known for some time, the myofibroblast has been however, there is no normal cellular counterpart: with I identified only comparatively recently.1•2 Partly on this or 2 exceptions,8•9 the myofibroblast is not found in account, it is a less well understood cell. In spite of this, untraumatised adult tissues (see refs 6,10,11). Myofi­ many lesions have been classified as myofibroblastic broblasts in granulation tissue and tumor stroma are (Table 1). For some of these lesions (especially nodular argued here as the nearest equivalent to the normal cell fasciitis, the fibromatoses-including the nodular pal­ counterpart with which to define and identify myofi­ mar variant, Dupuytren's disease-and inflammatory broblastic lesions. Such myofibroblasts (referred to myofibroblastic tumor), there is widespread agreement here as "reactive" myofibroblasts) are recognized by on their myofibroblastic nature. For other lesions, some characteristic light and electron microscopy features. authors have found the claims for myofibroblastic dif­ They are spindled cells with fusiform nuclei, set in a ferentiation unconvincing (for example, the myofibro­ more or less collagenized matrix, with a rather ill­ blastomas3-5). Such controver�ies have arisen partly defined cytoplasm, which is paler and less fibrillar than because of the difficulty in reaching an agreed defini­ the usually brightly eosinophilic cytoplasm typical of tion for the myofibroblast. Because the myofibroblast smooth-muscle cells12 (Fig 1). They stain for a-smooth­ was initially defined in ultrastructural terms, 1•2 electron muscle actin, fibronectin, and vimentin. Lesional myo­ microscopy has been central to identifying the myofi­ broblast and diagnosing myofibroblastic tumors and also express some desmin staining, a finding lesions. More recently, the convenience of a definition that requires some comment. Tumors cannot be ex­ based on light microscopy and immunohistochemistry pected to mirror exactly their putative normalcell coun­ has been argued.6•7 Within this context, this review terparts12; the genetic abnormalities that characterize deals with the main applications and controversies of electron microscopy in diagnosing myofibroblastic le­ sions. From the Department of Histopathology, Christie Hospital Na­ tional Health Service Trust, Manchester, United Kingdom. Address reprint requests to Brian Eyden, PhD, Head of Diagnostic THE DEFINITION OF THE MYOFIBROBlAST Electron Microscopy, Department of Histopathology, Christie Hos­ AS THE BASIS FOR IDENTIFYING pital NHS Trust, Manchester M20 4BX, United Kingdom; e-mail: MYOFIBROBlASTIC LESIONS [email protected]. Copyright 2003, Elsevier Inc. All rights reserved. Diagnosing tumors partly consists of identifying fea­ 0740-2570103/2001-0003$30.0010 tures observed in a putative normal cellular counterpart: doi: JO.l053/sdia.2003.50004

Seminars in Diagnostic Pathology, VOL 20, NO 1 (FEBRUARY), 2003: pp 13-24 13 14 BRIAN EYDEN

Table 1. Myofibroblastic Lesions: Immunohistochemical and Ultrastructual Features

References A' D RER MF FN FF/FNXIL Comments

Keloid 26 + + + FNX SMA 27 + + + + FNX SMA Hypertrophic Scar 28 + + + FNX SMA 27 + + + + FNX SMA Nodular fasciitis 29 + + FF FF2 (Fig 6) 24 FNX Proliferative fasciitis 30 + + FF FF2 (Fig 2) Proliferative myositis 31 + + + + SMA, D in V4 cases Myositis ificans oss 32 + + FF FF2 (Figs 16, 17} Fibromatosis 33 + + FF Ff2 (Fig4), IB 34 + + + FF MSA, SMA, Ffi2 (Fig 6), IB 35 + + L 36 + + FNX FF2 {eg. Fig 4)3 Inflammatory myofibrobtastic tumor 37 + L L4 38 + + + L AMA, footnote 4 39 + + + FF FF2 (Fig 3) 40 + + + FF SMA, MSA, CK, FF2 (Fig 3). 41 + + FNX Figs 5, 6, 7a 24 FNX Proliferative funiculitis 42 + + SMA, MSA, CK Post-operative spindle cell nodule 43 + + + + FF SMA,CK This paper + + FF Atypical decubital fibroplasia 44 + MSA Fibroma of tendon sheath 24 FF Myofibrosarcoma 45 + + FNX FNX (Fig 4), Ff2 17 + + + FF SMA 46 + + + + FNX SMA, (Figs 9-11) 47 + + + + FNX SMA (Figs 6,7) 6 + + + + + L SMA, MSA, L4 48 + + + + FNX SMA 7 + SMA 49 + + + ?FF SMA,MSA 50 + + + + FNX- SMA, MSA Malignant fibrous histiocytoma 51 + + FF Ff2 (Fig 6b) 52 + FF CK and tonofibrils, FF (Figs 4,5) 53 + (+) + FNX ULTRASTRUCTURE OF MYOFIBROBLASTIC LESIONS 15

Table 1. (Cont'd) Myoflbroblastic Lesions: Immunohistochemical and Ultrastructual Features

References D REA MF FN FF/FNX/L Comments

Myofibroblastoma 54 + + + Footnote 5 55 + + 56 + L Ap4 57 + + + L SMA, AP4 58 + + + FF 3 + + + L SMA 4 + + + + L SMA 48 + + SMA, D + 3/5 Cases 59 + + + FF,L 60 + + + + L SMA, D + 5/5 Cases, AP4•6 5 + + + L SMA, MSA, no focal densities Angiomyofibroblastoma 61 + + L D+ 10/10 cases, little rEA Superficial cervico-vaginal myofibroblastoma 62 + + D + 13/13 cases, SMA, MSA Congenital/infantile fibrosarcoma 63 + Dermatofibroma 64 + + + L SMA, MSA, AP4 Dermatofibrosarcoma protuberans 65 + D-ve in 15/15 cases 66 + + + SMA Atypical fibroxanthoma 65 + D-ve in 12/12 cases Dermatomyofibroma 67 + + + L MSA Uterine plexiform tumor 68 + L AP Paratesticular plexiform tumor 69 + + + MSA, AP4·7 Leiomyomatosis peritonealis disseminate 70 + + L Massive ovarian edema 71 + + Pseudoangiomatous stromal hyperplasia 72 + + SMA Fibrous hamartoma of infancy 73 + + 74 + + + SMA Elastofibroma 75 Footnote 8, abundant IF Nasopharyngeal angiofibroma 76 + + + SMA

Abbreviations: A, actin; SMA, a-smooth-muscle actin; MSA, muscle specific actin (using HHF35); AMA, anti-muscle actin (equivalent to MSA); D, desmin; CK, cytokeratin; FN, fibronectin; IF, intermediate filaments seen by electron microscopy and not specified as to type; rEA, rough endoplasmic reticulum; MF, myofilaments with focal densities; FF, fibronectinfibril; FNX, fibronexus; L, lamina (Kbasement membrane"); AP, attachment plaque; 18, inclusion body. 2. Not referred to as such or not referred to at aU and variously described as basement membrane-like material, microtendon, filamentous extracellular material, basal lamina-like material or hemi-desmosome-like material. 3. Dupuytren's disease 4. Indicates smooth-muscle differentiation 5. lntranodal hemorrhagic spindle-cell tumor with amianthoid fibers 6. Smooth-muscle differentiation in a minority of cells 7. Referred to as myofibroblastoma 8. Fibroblastic lesion 16 BRIAN EYDEN

Fig 1. Light microscope morphology of myofibroblastic lesions. (A) Nodular fasciitis (same case as in Figure 5). (B) Myofibrosarcoma infiltrating fat (same case as in Figure 8). Arrowheads indicate plump spindled tumor cells. (A & B) H&E, x300. the neoplastic process can explain loss of anticipated broblast derives from the idea that it arises from a features, as well as the appearance of unexpected ones precursor cell that undergoes a series of differentiation (such findings from the field of immunohistochemistry steps, and that for some pathologists it has been diffi­ are often referred to as anomalous or aberrant). cult to define the precise location of the myofibroblast Desmin staining in myofibroblastic lesions is interest­ within this dynamic process. The widely accepted ing because granulation tissue and tumor stroma myo­ working hypothesis is of a quiescent fibroblast (Fig 2) fibroblasts express very little of this intermediate fila­ which, on activation, synthesises abundant rough endo­ ment13·14. Desmin, therefore, should not be considered plasmic reticulum cisternae (rER) for matrix produc­ partof the primarydefinition of the myofibroblast,even tion, then switches on genes for smooth-muscle actin though it is recognised in some lesions. It is compatible myofilaments for contractility.15 These 2 ultrastructural with myofibroblastic differentiation, but arguably it features especially (Fig 3) (but also, for example, gap would be inappropriate to use it to support a myofibro­ junctions and collagen secretion granules16-19) are blastic interpretation-justas, by analogy, the fair num­ widely understood and accepted as markers for the ber of malignant melanomas which are positive for myofibroblast, but this cell also has a further distinctive cytokeratin would not justify making cytokeratin a pri­ feature, at the cell surface - the fibronexus, which is mary marker for melanoma on the same level of im­ less well recognized (Fig 3). portance as SlOO protein or HMB45 (see below for The first detailed ultrastructural descriptions of the further discussion) myofibroblast1•2 not only described rER and myofila­ One of the difficulties in understanding the myofi- ments but also a component on the cell surface desig- ULTRASTRUCTURE OF MYOFIBROBLASTIC LESIONS 17

1 myofibroblast and identifying myofibroblastic differen­ Quiescent Fibroblast tiation in tumors and tumor-like lesions. The features conform to stage 4 in the fibroblast-myofibroblast trans­ formation (Fig 2) and collectively amount to arguably the highest level of myofibroblastic differentiation pos­ sible ("fully differentiated" myofibroblasts). The extent to which these features arefound in the wide variety of lesions described as myofibroblastic is documented in Table 1.

SPECTRUM OF DIFFERENTIATION IN MYOFIBROBLASTIC TUMORS AND TUMOR­ LIKE LESIONS: UNDER-IDENTIFICATION OF THE FIBRONEXUS

A number of broad conclusions, as well as some areas of controversy, arise from the data in Table 1. Using the ultrastructural criteria of fibronexus junc­ tions, prominent rER and modestly developed periph­ eral myofilaments as representing the highest level of myofibroblastic differentiation {in a consonant histo­ logical and immunophenotypic picture), a spectrum of myofibroblastic differentiation is evident. Those lesions containing the most highly differentiated myofibro­ Fig 2. Diagrammatic representation of the fibroblast­ blasts are hypertrophic scar, Dupuytren's disease {Fig myofibroblast transformation. Four stages are Identifiable 4} and other fibromatoses, nodular and proliferative ultrastructurally. An activated fibroblast showing periph­ eral myofllaments but not fibronexuses ("myoid" fibro­ fasciitis (Fig 5), and inflammatory myofibroblastic tu­ blast) is a theoretical precursor to the fully differentiated mor. Those lesions with lesser but still distinct myofi­ myofibroblast (stage 4). Both stages 3 and 4 would be broblastic differentiation include keloid, fibroma of ten­ compatiblewith the light microscope criterionfor the myo­ fibroblast of a spindled-cell positive for smooth-muscle don sheath and post-operative spindle-cell tumor (Fig actin. rEA, rough endoplasmic reticulum; SLD, subplas­ 6). These less well-differentiated lesions depend to a malemmal linear density; CSG, collagen secretion granule. certain extent for their identification on the ability to identify fibronectin fibrils on their own, since the fi­ nated as basement membrane-fike material. Further bronexus-associated bundles may be ab­ study showed that this differed from basal lamina sent or sparse. In this respect, it is important to be able ("basement membrane") and represented the external to distinguish the fibronectin fibril of the myofibroblast component of a cell-surface specialization called the from the lamina of smooth-muscle differentiation. fibronexus or fibronexus junction.2•20-24 At the fibro­ Compared with lamina, the fibronectin fibriltends to be nexus, intracellular myofilaments and extracellular fi­ more densely staining, straighter in profile, with a finely bronectin filaments (these forming the fibronectin filamentous substructure, often a co-linear relationship fibri/23·24) converge to form a cell-to-matrix adhesive with nearby intracellular myofilaments, and is often device. Fibronexus junctions have been promoted as an seen diverging away from the cell surface(Figs 3B and important myofibroblastic feature for two reasons: they 5; and refs 24 and 25). formed part (albeit under a different terminology) of It is worth emphasising that the fibronexus and the the original definition of the granulation tissue myofi­ fibronectin fibril are under-interpreted structures in that broblast, 1 and they are consistently and sometimes several papers exist (Table I) where these structuresare exuberantly expressed in tumor stroma myofibro­ illustrated but either not mentioned or are misinter­ blasts.2,23·25 preted as lamina or basement-membrane-like material These features - rER, peripheral myofilaments and of indeterminate significance.. Examples include nodu­ fibronexus junctions - in an appropriatehistology and lar and proliferative fasciitis,29.3° myositis ossificans,32 immunophenotype, constitute one basis fordefining the as well as some fibromatoses,34 and inflammatory myo- 18 BRIAN EYDEN

Fig 3. Myofibroblasts from tumor stroma (keratinizing squamous cell carcinoma, skin of neck in an 85 year old man). (A) Abundant rEA, peripheralmyofilaments, and cell surface fibronectin fibrils. Note abundant intercellularcol· lagen. x5,000. (Reprinted with permission.24) (B) Detail of flbronectin fibril and myofllaments at a fibronexus. Note that the fibronectin fibril has high staining density, co· linearity with myofilaments and filamentous substructure. x12,000. {Reprinted with permission.25) Electron micro· graphs. Abbreviations and labelling {Figs 3-8): *", rough endoplasmic reticulum; arrow, myofilaments; arrowhead, fibronectin fibril; L, lamina; Go, Golgi apparatus; N, nu· cleus; co, collagen. ULTRASTRUCTURE OF MYOFIBROBLASTIC LESIONS 19

Fig 4. Dupuytren's disease. {A) Spindled cells with rER, a Golgi apparatus, myofllaments and flbronectin fibrils, amid abundant collagen. x5,900. (B) Fibronectin fibril and myofilaments at a fibronexus. x30,000. (C) Golgi apparatus with collagen secretion granules (small arrows). x30,000. (Specimen courtesy Professor Hartwig Kosmehl and Dr Alex Berndt, Jena, Germany).

fibroblastic tumors.39 In short, the incidence of the SMOOTH-MUSCLE DIFFERENTIATION IN SO fibronexus or fibronectin-fibril is higher than is recog­ CALLED MYOFIBROBLASTIC LESIONS nised, and true myofibroblastic differentiation is there­ fore arguably more widespread than perceived. Just as some so called myofibroblastic lesions are Often, these myofibroblastic lesions are described as more properly to be regarded as fibroblastic, there are consisting of myofibroblastic and fibroblastic cells to compelling reasons on the grounds of ultrastructure reflect observations that not all cells have the ultrastruc­ and/or strong desmin staining to consider certain other tural features of myofibroblasts. However, some le­ lesions as expressing a kind of smooth-muscle (non­ sions, such as elastofibroma,75 appear to be essentially myofibroblastic) differentiation. While many of the ul­ fibroblastic, in which abundant intermediate filaments trastructurally examined fibromatoses, as already indi­ have been misinterpreted as actin (ie, muscle) fila­ cated, exhibit fibronexus junctions and so are fully ments. myofibroblastic, a few others have the lamina indica- BRIAN EYDEN

since some fibromatoses show smooth-muscle ultra­ structure. 35 In identifying smooth-muscle differentiation fine structurally, attachment plaques with overlying lamina (sometimes with intervening plasmalemma! caveolae) constitute an important smooth-muscle feature distinct from the fibronexus of the myofibroblast.25 These smooth-muscle surface features have been seen in a number of myofibroblastomas (Fig 7).3-5.56,57 By con­ trast, there is only one study where there is a suspicion of a fibronectin fibril.ss Many of the tumors referred to as myofibroblastomas may, therefore, be exhibiting a kind of (nonmyofibroblastic) smooth-muscle differen­ tiation. This argument, especially given the strong desmin staining in some cases, applies also to the angiomyofibroblastomas. Here, the cytoplasm contains little of the rER expected of the myofibroblast, no myofilaments, the fibronexus has never been identified, and one study has even shown smooth-muscle lam­ ina.6t

THE LIGHT MICROSCOPY DEFINITION OF THE MYOFIBROBLAST

Other so called myofibroblastic lesions that are ar­ guably showing a form of smooth-muscle differentia­ Fig 5. Lesional myofibroblast in a nodular1asciltiS (su­ prapubic mass in a 66-year·old woman) showing myofila­ tion include paratesticular plexiform tumor,69leiomyo­ ments and a straight flbronectin flbrU projecting into ex· matosispe ritonealis disseminata70 (which appears to be tracellular space. x8,000. (Reprinted with pennission.24) expressing a synthetic or matrigenic smooth-muscle phenotype to judge by the prominent rER), and massive tive of smooth-muscle cells35 or are strongly desmin­ ovarian edema.71 In certain other lesions-dermatofi­ positive,77 a feature suggesting smooth-muscle rather broma,64 Dermatofibrosarcoma protuberans,65,66 atypi­ than myofibroblastic differentiation. It is interesting cal fibroxanthoma,65 dermatomyofibroma,67 pseudoan­ that in Hasegawa's paper, Dupuytren's disease and giomatous stromal hyperplasia72 - the evidence for nodular fasciitis, which have· well-defined myofibro­ myofibroblastic differentiation is based either on the blastic credentials on the basis of fibronexus junc­ less strict ultrastructural combination of rER with myo­ tions/4·36 are far less desmin-reactive than extra-ab­ filaments (without fibronexuses- stage 3 in Fig 2), or dominal desmoid. In addition, in separate studies of 103 the light microscopy definition encompassing spindle­ examples of nodular fasciitis, none were reactive for cell morphology and immunstaining for a-smooth­ desmin. 78·79 These findings further emphasise the "true" muscle actin and/or desmin. or "complete" myofibroblastic phenotype in nodular This light microscopy definition merits attention be­ fasciitis. cause of the acknowledged decline in the use of elec­ Desmin has been used as a marker for myofibroblast tron microscopy.6·7 It remains an argument held by lesions13 but it should be remembered that this inter­ some authorities, however, that the light microscopy mediate filament is not well expressed in reactive ("nor­ definition is more imprecise and open to greater inter­ mal") myofibroblasts, although it is prominent in a pretational uncertainty in the sense that a variety of number of myofibroblastic lesions such as the fibroma­ spindled cells can express a-smooth-muscle actin - toses. It has to beconside red whether in these situations smooth-muscle cells, myofibroblasts, , myoep­ one should consider desmin staining as a further marker ithelial cells, undergoing mesenchymal for myofibroblastic differentiation, or whether it should transformation.25 To illustrate this point further, elec­ be more appropriate to think of it as indicating an tron microscopy can distinguish, for example, a myo­ element of smooth-muscle differentiation, especially fibroblastfrom a matrigenic smooth-muscle cell, ie, one ULTRASTRUCTURE OF MYOFIBROBLASTIC LESIONS 21

Fig 6. Postoperative spindle-cell nodule (vulva, 33year old). (A) Lesional myoflbroblasts containing variable num­ bers of rERcisternae, and very modestly developed myo­ fllaments. One cell process has a flbronectin fibril. Ma, . x7,100.(B) Detail of fibronectin fibril from A. x32,000. (C) Nearby endothelium to show different ultra­ structural appearance of lamina. x32,000. which has transdifferentiated from the contractile to the MYOFIBROBLASTIC MALIGNANCIES synthetic phenotype, and which, in spite of containing abundant rER and relatively few peripheral myofila­ The early studies on tumor stroma and granulation ments like a myofibroblast, differs from it by its tissue emphasised the reactive nature of the myofibro­ smooth-muscle cell surface.25 The ultrastructural defi­ blast.81 Increasingly in recent years, however, spindle­ nition is useful, therefore, in situations of diagnostic cell malignancies with myofibroblastic differentiation uncertainty associated with spindle-cell tumors positive have been documented.17•4547•53•82 Figures 1 and 8 il­ for smooth-muscle actin and or desmin. An additional lustrate an unambiguous myofibroblastic sarcoma area of utility is for intra-abdominal spindle-cell lesions (myofibrosarcoma) where the patient died of metastatic positive for c-kit, which might be true gastro-intestinal disease.n Importantly, the bulk of the tumor consisted stromal tumors or myofibroblastic.80 of myofibroblasts bearing many fibronectin fibrils, ex- BRIAN EYDEN eluding the argument that the myofibroblastic differen­ tiation in such tumors was due to reactive cells. Any poorlydiffe rentiated spindle-cell sarcoma couktbenefit from ultrastructural input on the basis that some may show evidence of myofibroblasticdifferentiatwn. 49 The complex and controversial entity, malignant fibrous histiocytoma, also contains variable numbers of myo­ fibroblasts as neoplastic elements. s2.S3•82

ACKNOWLEDGEMENTS

My sincere thanks go to Professor Hartwig Kosmehl (Erfurt, Gennany} and DrAlexander Berndt (Jena, Germany) forsupplying a resin block of Dupuytren's disease; Professors F Hernandez, JV Johannessen, JM Nesland and B Baccetti, for permission to reprint figures; Dr John Coyne (Manchester, UK) for providing .clinical

Fig 8. Myoflbrosarcoma (case 217- subcutaneous tis­ sue of axilla in a 77-year-old woman). (A) Note rEA, myo­ filaments and multiplefibronectln fibrils. x4, 100.(Reprinted with permission.11) (B) Detail of dense, straightfibronectin fibril. x30,000.

details of the case of postoperative spindle cell nodule; V arsha Mistry for word-processing assistance; and Paul Chantry for drawing Fig­ ure 2.

REFERENCES

1. Gabbiani G, Ryan GB. Majno G: Presence of modified fibro­ blasts in granulation tissue and their possible role in wound contrac­ tion. Experientia 27: 549-550, 1971 2. Ryan GB, Cliff WJ, Gabbiani G, et a!: Myofibroblasts in human Fig 7. Myofibroblastoma fromthe inguinal region of a granulation tissue. Hum Pathol 5:55-67, 1974 53-year-old man.s Well-defined focal lamina. x32,000. (Re­ 3. Eyden BP, Harris M, Greywoode GI, et a!: Intranodal myofi­ printed with permission from the Journal of Submicro­ broblastoma: report of a case. Ultrastruct Pathol 20: 79-88, 1996 scopicCytology and Pathology. 5) 4. Eyden BP, Shanks JH, Ioachim E, et al: Myofibroblastoma of ULTRASTRUCTURE OF MYOFIBROBLASTIC LESIONS 23

breast: evidence favoring smooth-muscle rather than myofibroblastic 24. Eyden B: The fibronexus in reactive and tumoral myofibro­ differentiation. Ultrastruct Pathol 23: 249-257, 1999 blasts: further characterisation by electron microscopy. Histol His­ 5. Eyden B. Chorneyko K: Intranodal myofibroblastoma: study of topathol 16: 57-70, 2001 a case suggesting smooth-muscle differentiation. J submicrosc Cytol 25. Eyden B: The myofibroblast: An assessment of controversial Pathol 33:157-163, 2001 issues and a definition useful in diagnosis and research. Ultrastruct 6. Mentzel T, Dry S, Katenkamp D, et al: Low-grade myofibro­ Pathol 25: 39-50, 2001 blastic sarcoma: analysis of 18 cases in the spectrum of myofibro­ 26. Shoukrey NM, Tabbara KF. Ultrastructural stndy of a corneal blastic tumors. Am J Surg Pathol 22:1228-1238, 1998 keloid. Eye 7: 379-387, 1993 7. Bisceglia M, Magro G: Low-grade myofibrosarcoma of the 27. Santucci M, Borgognoni L, Reali UM, et al: and salivary gland. Am I Surg Path 23: 1435-1436, 1999 hypertrophic of Caucasians show distinctive morphologic and 8. Holstein AF, Maekawa M, Nagano T, et al: Myofibroblasts in immunophenotypic profiles. Virchows Arch 438: 457-463, 2001 the of human seminiferous tubules are dynamic struc­ 28. Ehrlich HP, Desmouliere A, Diegelmann RF, et al: Morpho­ 59: 109-125, tures of heterogeneous phenotype. Arch Histol Cyto1 logical and immunochemical differences between keloid and hyper­ 1996 trophic scar. Am I Pathol 145: 105-113, 1994 9. Garant PR: Collagen resorption by fibroblasts. A theory of 29. Wirman JA: Nodular fasciitis, a lesion of myofibroblasts: An fibroblast maintenance of the periodontal . I Periodont 47: ultrastructural study. Cancer 38: 2378-2389, 1976 380-390, 1976 30. Craver JL, McDivitt RW: Proliferative fasciitis. Ultrastruc­ 10. Yang P, Hirose T, Seki K, et al: Myofibroblastic tumor of soft tural study of two cases. Arch Pathol Lab Med 105: 542-545, 1981 tissue displaying desmin-positive and actin-negative irnmunopheno­ 31. El-Jabbour JN, Bennett MH, Burke MM, et al: Proliferative types. Pathol Int 46: 6%-703, 1996 myositis. An immunohistochemical and ultrastructural study. Am J ll. Hisaoka M, Hashiomoto H, Daimaru Y: Intranodal palisaded Surg Pathol 15: 654-659, 1991 myofibroblastoma with so-called amianthoid fibers: a report of two 32. Povysil C, Matejovsky Z: Ultrastructural evidence of myofi­ cases with a review of the literature. Pathol 1nt 48: 307-312, 1998 broblasts in pseudomalignant myositis ossificans. Virchows Arch A 12. Mentzel T: Myofibroblastic sarcomas: a brief review of sar­ Path Anal and Histol 381: 189-203, 1979 comas showing a myofibroblastic line of differentiation and discus­ 33. Bhawan J, Bacchetta C, Majno G: A myofibroblastic tumor. sion of the differential diagnosis. Curr Diagn Pathol 7: 17-24, 2001 Infantile digital fibroma (recurrent digital fibrous tumor of child­ 13. Skalli 0, Schiirch W, Seemayer T, et al: Myofibroblasts from hood). Am I Pathol 94: 19-36, 1979 diverse pathologic settings are heterogeneous in their content of actin 34. Viale G, Doglioni C, Iuzzolino P, et al: Infantile digital fibro­ isoforms and intermediate filament proteins. Lab Invest 60: 275-285, matosis-like tumour (inclusion body fibromatosis) of adulthood: re­ 1989. port of two cases with ultrastructural and immunohistochemical find­ 14. Truong LD, Rangdaeng S, Cagle P, et al: Thediagnostic utility ings. Histopathology 12: 415-424, 1988 of desrnin. A study of 584 cases and review of the literature. Am J 35. Takagi M, Yamamoto H, Mega H, et al: Heterogeneity in the Clin Pathol 93: 305-314, 1990 gingival fibromatoses. Cancer 68: 2202-2212, 1991 15. Welch MP, Odland GF, Clark RAF: Temporal relationships of 36. Tomasek JJ, Schultz RJ, Haaksma CJ: ­ F-actin bundle formation, collagen and fibronectin matrix assembly, cytoskeletal connections at the surface of the specialized contractile and fibronectin receptor expression to wound contraction. J Cell Bioi fibroblast (myofibroblast) in Dupuytren disease. J Joint Surg 110: 133-145, 1990 Am 69: 1400-1407, 1987 16. Eyden BP: Collagen secretion granules in reactive stromal 37. Buell R, Wang N-S, Seemayer TA, et al: Endobronchial myofibroblasts, with preliminary observations on their occurrence in plasma cell granuloma (xanthomatous pseudotumor). Hum Path 7: spindle cell tumours. Virchows Arch [AJ415: 437-445, 1989. 411-426,1976 17. Eyden BP, Baneljee SS, Harris M, et al: A study of spindle cell 38. Gilks CB, Taylor GP, Clement PB: Inflammatory pseudotu­ sarcomas showing myofibroblastic differentiation. llitrastruct Pathol 6: 275-286, 1987 15: 367-378, 1991. mor of the uterus. IntI Gynecol Pathol 39. 18. Eyden B: Organelles in human tumor diagnosis: an ultrastruc­ Chou P, Gonzalez-Crussi F, Cole R, et al: Plasma cell gran­ 62: 1409-1413, 1988 tural atlas. New York and Tokyo,Igaku-Shoin, pp 68-70, 1996. uloma of the heart. Cancer 40. 19. Schiirch W, Seemayer TA, Gabbiani G: The myofibroblast: A Lundgren L, Aldenborg F, Angervall L, et al: Pseudomalig­ quarter century after its discovery. Am I Surg Pathol 22: 141-147, nant spindle cell proliferations of the urinary bladder. Hum Pathol25: 1998. 181-191, 1994 20. Baur PS, Barratt G, Linares HA, et al: Wound contractions, 41. Biselli R, Boldrini R, Ferlini C, et al: Myofibroblastic tu­ scar and rnyofibroblasts: a classical case study. J Trauma mours: neoplasias with divergent behaviour. Ultrastructural and flow 18: 8-22, 1978 cytometric analysis. Path Res Pract 195: 619-632, 1999 21. Baur PS Jr, Parks DH: The myofibroblast anchoring strand­ 42. Hollowood K, Fletcher CDM: Pseudosarcomatous myofibro­ the fibronectin connection in and the possible loci of blastic proliferations of the spermatic cord ("proliferative funicula­ collagen fibril assembly. I Trauma 23: 853-862, 1983 tis"). Histologic and immunohistochemical analysis of a distinctive 22. Singer II, Kawka DW, Kazazis DM, et al: In vivo codistribu­ entity. Am J Surg Pathol 16: 448-454, 1992 tion of fibronectin and actin fibers in granulation tissue. Immunoflu­ 43. Papadimitriou JC, Drachenberg CB: Posttraumatic spindle cell orescence and electron microscope stndies at the myofibroblast sur­ nodules. Immunohistochemical and ultrastructural study of two le­ face. J Cell Bioi 98: 2091-2106, 1984 sions. Arch Pathol Lab Med 118: 709-711, 1994 23. Eyden BP: Brief review of the fibronexus and its significance 44. Montgomery EA, Meis JM, Mitchell MS, et al: Atypical for myofibroblastic differentiation and tumor diagnosis. Ultrastruct decubital fibroplasia. A distinctive fibroblastic pseudoturnor occur­ Pathol 17: 611-622, 1993 ring in debilitated patients. Am J Surg Pathol 16: 708-715, 1992 BR.IAN EYDEN

45. Vasudev KS, Marris M: A sarcoma of myofibroblasts. Arch 64. Zeiger BWH, Zeiger BG, Rappersberger K: Prominent myo­ Pathol Lab Med 102: 185-188,1978 fibroblastic differentiation. A pitfall in the diagnosis of dermatofi­ 46. Eyden BP, Christensen L, Tagore V, et al: Myofibrosarcoma broma. Am J Dermatopath 19: 138-146, 1997 of subcutaneous of the cheek. 1 submicrosc Cytol Patbol 65. Ma CK, Zarbo RJ, Gown AM: Immunohistochemical charac­ 24: 307-313, 1992 terization of atypicalfibroxanthoma anddermatofibrosarcoma protu­ 47. Taccagni G, Rovere E, Masullo M, et al: Myofibrosarcomaof berans. Am .J Ciin Pathol97: 478-483, 1992 the breast: Review of the literature on myofibroblastic tumors and 66. Dominguez-Malag6n HR, Ordoiiez NG, Mackay B: Dermat­ criteria for definingmyofibroblastic differentiation. Am J Surg Pathol ofibrosarcoma ptotuberans: ultrastructural and immunocytochemical 21: 489-496, 1997 observations. Ultrastruct Pathol 19: 281-289, 1995 48. Gocht A, Bosmuller HC, Bassler R, et al: Breast tumors with 67. Colome Ml, Sanchez RL: Dermatomyofibroma: report of two myofibroblastic differentiation: clinico-pathological observations in cases. J Cutan Pathol21: 371-376, 1994 myofibroblastoma and myofibrosarcoma. Pathol Res Pract 195: 1-10, 68. Fisher ER. Paulson JD, Gregorio RM: The myofibroblastic 1999 nature of the uterine plexiform tumor. Arch Pathol Lab Med 102: 49. Bisceglia M, Tricarico N, Minnenna P, et al. Myofibrosarcoma 477-480, 1978 of the upper jawbones: a clinicopathologic and ultrastructural study 69. Busmanis 1: Paratesticular plexiform tumour of myofibroblas­ of two cases. Ultrastruct Pathol 25: 385-397, 2001 tic origin. Histopathology18: 178-180, 1991 50. Montgomery E, Goldblum JR, Fisher C: Myofibrosarcoma. A 70. Pieslor PC, Orenstein JM, Hogan DL, et a!: Ultrastructure of clinicopathologic study. Am J Surg Pathol 25: 219-228, 2001 myofibroblasts and decidualized cells in leiomyomatosis peritonealis 51. Fukuda T, Tsuneyoshi M, Enjoji M: Malignant fibrous histi­ disseminata. Am J Clin Pathol: 72: 875-882, 1979 ocytoma of soft parts: An ultrastructural quantitative study. Ultra­ 71. Roth LM, Deaton RL, Sternberg WH: Massive ovarian edema. struct Pathol 12: 117-129, 1988 A clinicopathologic study of five cases including ultrastructural ob­ 52. Meek WD: Fibronexus in a case of malignant fibrous histio­ servations and review of the literature. Am J Surg Pathol 3: 11-21, cytoma. Ultrastruct Pathol 18: 365-370, 1994 1979 53. Watanabe K, Ogura G: Fibronexus in "malignant fibrous his­ 72. Powell CM, Cranor ML, Rosen PP: Pseudoangiomatous stro­ tiocytoma" of the bone: A case report of pleomorphic myofibrosar­ mal hyperplasia (PASH). A mammary stromal tumor with myofibro­ coma. Ultrastruct Pathol 26: 47-51, 2002 blastic differentiation. Am J Surg Pathol 19: 270-277, 1995 54. Suster S, Rosai J: lntranodal hemorrhagic spindle-cell tumor 73. AbererA, Mainitz M, EntacherU, et al: Fibrous hamartoma of with "amianthoid" fibers. Report of six cases of a distinctive mesen­ infancy - infantile subcutaneous myofibroblastoma. Derrnatologica chymal neoplasm of the inguinal region that simulates Kaposi's 176: 46-51, 1988 sarcoma. Am J Surg Pathol 13: 347-357, 1989 74. Groisman G, Lichtig C: Fibrous hamartoma of infancy: An 55. Weiss SW, Gnepp DR, Bratthauer GL: Palisaded myofibro­ immunohistochemical and ultrastructural study. Hum Pathol 22: 914- blastoma. A benign mesenchymal tumor of lymph node. Am J Surg 918, 1991 Pathol 13: 341-346, 1989 75. Ramos CV, Gillespie W, Narconis RJ: Elastofibroma. Arch 56. Michal M, ChlumskJIA, Ska!ova A, et al: Palisaded intranodal Patbol Lab Med 102: 538-540, 1978 myofibroblastoma. Electron microscopic study. Zentralbl f Pathol 76. Beham A, Kainz J, Stammberger H, et al: Immunohistochem­ 139: 81-88, 1993 ical and electron microscopical characterization of stromal cells in 57. Tanda F, Massarelli G, Cossu A, et al: Primary spindle cell tumor of lymph node with "amianthoid" fibers: a histological, im­ nasopharyngeal angiofibromas. Eur Arch Otolaryngol 254: 196-199, 1997 munohistochemical and ultrastructural study. Ultrastruct Pathol 17: 195-205, 1993 77. Hasegawa T, Hirose T, Kudo E, et al: Cytoskeletal character­ 58. Padberg BC, Holzhausen HI, Weinland G, et al: Primlirer istics of myofibroblasts in benign neoplastic and reactive fibroblastic hliruorrhagischerspindelzelltumor des lymphknotens mit amianthoid­ lesions. Virchows Arch A Pathol Anat 416: 375-382, 1990 strukturen.Fallbericht und literaturiibersicht. Pathologe 15: 176-180, 78. Montgomery EA, Meis JM: Nodular fasciitis. Its morphologic 1994 spectrum and immunohistochemical profile. Am J Surg Pathol 15: 59. CarneiroF, Gon�alves V, Sobrinho SimOes M: Myofibroblas­ 942-948, 1991 toma of the meninges. Ultrastruct Pathol 13: 599-605, 1989 79. Thompson LDR, Fanburg-Smith JC, Wenig BM: Nodular 60. Herrera G, Johnson WW, Lockard VG, et al: Soft tissue fasciitis of the external ear region: a clinicopathologic study of 50 myofibroblastomas. Mod Pathol 4: 571-577, 1991 cases. Ann Diagn Pathol 5: 191-198, 2001 61. Fletcher CDM, Tsang WYW, Fisher C, et al: Angiomyofibro­ 80. Y antiss RK, Spiro IJ, Compton CC, et al: Gastrointestinal blastoma of the vulva. A benign neoplasm distinct from aggressive stromal tumor versus intra-abdominal fibromatosis of the bowel wall. angiomyxoma. Am J Surg Pathol 16: 373-382, 1992 A clinically important differential diagnosis. Am J Surg Pathol 24: 62. Laskin WB,Fetsch J, Tavassoli FA: Superficial cervicovaginal 947-957, 2000 myofibroblastoma: fourteen cases of a distinctive meseiiChymal tu­ 81. Seemayer TA, Schiirch W, Lagace R: The myofibroblast and mor arising from the specialized subepithelial stroma of the lower defense against neoplasia: a hypothesis. Surv immunol Res l: 268- female genital tract. Hum Pathol 32: 715-725,2001 273,1982 63. Argyle JC, Tomliuson GE, Stewart D, et al: Ultrastructural, 82. Montgomery E, Fisher C: Myofibroblastic differentiation in immunocytochemical, and cytogenetic characterization of a large malignant fibrous histiocytoma (pleomorphic myofibrosarcoma): A congenital fibrosarcoma. Arch Pathol Lab Med 116: 972-975, 1992 clinicopathological study. Histopathology 38: 499-509, 2001