Possible involvement of inflammatory/reparative ... - Springer Link

4 downloads 0 Views 2MB Size Report
Nov 27, 2015 - Uterine leiomyomas (fibroids) are extremely common benign neoplasms. ... cess leading to the formation of a uterine fibroid with exagger-.
Cell Tissue Res (2016) 364:415–427 DOI 10.1007/s00441-015-2324-3

REGULAR ARTICLE

Possible involvement of inflammatory/reparative processes in the development of uterine fibroids Olga Protic 1 & Paolo Toti 2 & Md Soriful Islam 1,3 & Rossella Occhini 2 & Stefano Raffaele Giannubilo 4 & William H. Catherino 5 & Saverio Cinti 1,6 & Felice Petraglia 7 & Andrea Ciavattini 4 & Mario Castellucci 1 & Boris Hinz 8 & Pasquapina Ciarmela 1,9

Received: 14 May 2015 / Accepted: 29 October 2015 / Published online: 27 November 2015 # Springer-Verlag Berlin Heidelberg 2015

Abstract Uterine leiomyomas are benign tumors in the smooth muscle layer of the uterus. The most common histological type is the Busual leiomyoma^, characterized by overexpression of ECM proteins, whereas the Bcellular type^ has higher cellular content. Our objective is to investigate the involvement of inflammatory and reparative processes in leiomyoma pathobiology. Using a morphological approach, we investigate the presence of inflammatory cells. Next, we determine the localization of the ECM, the presence/absence of fibrotic cells via α-sma and desmin and the immunohistochemical profile of the mesenchymal cells with respect to CD34. Finally, we explore the effect of inflammatory mediators (TNF-α, IL-1β, IL-6, IL-15, GM-CSF and IFN-γ) on pro-fibrotic factor activin A mRNA expression in vitro. Higher numbers of macrophages were found inside and close to leiomyomas as compared to the more distant myometrium. Cellular leiomyomas showed more macrophages and mast cells than the Busual type^. Inside the fibroid tissue, we found cells positive for α-sma, but negative for desmin and a large amount of collagen surrounding the nodule, suggestive of

myofibroblasts producing ECM. In the myometrium and leiomyomas of the “usual type”, we identified numerous CD34+ fibroblasts, which are known to give rise to myofibroblasts upon loss of CD34 expression. In leiomyomas of the “cellular type”, stromal fibroblasts were CD34-negative. Finally, we found that TNF-α increased activin A mRNA in myometrial and leiomyoma cells. In conclusion, this study demonstrates the presence of inflammatory cells in uterine leiomyomas, which may contribute to excessive ECM production, tissue remodeling and leiomyoma growth. Keywords Usual leiomyoma . Cellular leiomyoma . Inflammation . ECM . Myofibroblast . CD34 . TNF-α . Activin A . Myometrium

Introduction Uterine leiomyomas (fibroids) are extremely common benign neoplasms. The incidence in women of reproductive age is

* Mario Castellucci [email protected]

4

Department of Clinical Science, Polytechnic University of Marche, 60131 Ancona, Italy

* Pasquapina Ciarmela [email protected]

5

Department of Obstetrics and Gynecology, Uniformed Services University of the Health Sciences, Bethesda, Maryland 20814, USA

6

Center of Obesity, Polytechnic University of Marche – United Hospitals, Ancona, Italy

7

Department of Molecular and Developmental Medicine, Obstetrics, and Gynecology, University of Siena, 53100 Siena, Italy

8

Laboratory of Tissue Repair and Regeneration, Matrix Dynamics Group, Faculty of Dentistry, University of Toronto, Toronto, Ontario M5S 3E2, Canada

9

Department of Information Engineering, Polytechnic University of Marche, Ancona, Italy

1

Department of Experimental and Clinical Medicine, Faculty of Medicine, Polytechnic University of Marche, via Tronto 10/a, 60020 Ancona, Italy

2

Department of Medical Biotechnology, University of Siena, 53100 Siena, Italy

3

Biotechnology and Microbiology Laboratory, Department of Botany, University of Rajshahi, Rajshahi 6205, Bangladesh

416

approximately 60 %, thus representing one of the major public health problems worldwide (Okolo 2008). The clinical symptoms include pelvic pain, discomfort and menstrual disorders, as well as infertility and depend on the position, size and number of leiomyomas (Buttram and Reiter 1981; EldarGeva et al. 1998). To date, there is no effective medical drug therapy against uterine leiomyomas. Some medical treatments have been shown to be effective in reducing leiomyoma volume; however, leiomyomas recur when patients exit the therapy (Friedman et al. 1989; Islam et al. 2013a). Hysterectomy, a permanent solution against leiomyomas, is a major abdominal surgical procedure with an increased risk of postoperative morbidity (Chegini 2010) and loss of female reproductive potential (Wallach and Vlahos 2004). Leiomyomas are often multiple and originate in the smooth muscle layer of the uterus. It has been proposed that each fibroid originates from a single transformed cell that proliferates, giving rise to a small leiomyoma that can expand to a size of more than 15 kg (Ezugwu et al. 2014; Ligon and Morton 2000; Mashal et al. 1994). Uterine leiomyomas are defined as benign neoplasms composed of smooth muscle cells with variable amounts of fibrous stroma. These can be subcategorized based on different histological features: (1) the “usual type”, most frequently diagnosed, (2) the “cellular type” that displays increased cellularity (Rosai 2011), (3) lipoleiomyomas, which are characterized by the presence of numerous adipocytes (Avritscher et al. 2001) and (4) rare “bizarre” variants (Ciarmela et al. 2012; Toledo and Oliva 2008). Leiomyomas are considered a fibrotic disorder (Chegini 2010; Leppert et al. 2006; Malik et al. 2010) displaying exaggerated and continuous wound healing triggered by tissue injury and characterized by excessive production of collagenous extracellular matrix (ECM) (Gelse et al. 2003). Numerous studies have shown that leiomyomas contain approximately 50 % more ECM, comprised of collagens, proteoglycans and fibronectin, than the surrounding myometrium (Fujita 1985,Arici and Sozen 2000; Norian et al. 2009; Stewart et al. 1994). Moreover, the structure and orientation of collagen fibrils were observed to be abnormal in uterine leiomyomas (Leppert et al. 2004). Despite the increased interest in the pathophysiology of leiomyomas in recent years (Islam et al. 2013b; Karmon et al. 2014), their etiology still remains unknown. In the present study, we investigate whether an inflammatory stimulus, acting on the myometrium, can trigger a reparative process leading to the formation of a uterine fibroid with exaggerate synthesis of ECM. Fibrotic responses develop upon recruitment of inflammatory cells to the site of injury and the activation of collagen producing fibroblasts (Kisseleva and Brenner 2008). These activated fibroblasts, also called myofibroblasts, regulate connective tissue remodeling by ECM synthesis/degradation and contraction of the surrounding tissue (Hinz et al. 2012). A hallmark of the contractile phenotype is the expression of αsmooth muscle actin (α-sma). Myofibroblasts, regardless of

Cell Tissue Res (2016) 364:415–427

their origin, are often activated by inflammatory signals and are responsible for restoring tissue homeostasis (Hinz 2007; Wynn 2008). However, inappropriate function of myofibroblasts can cause pathological fibrosis (Kisseleva and Brenner 2008). It has been suggested that smooth muscle cells of a leiomyoma can attain a myofibroblastic phenotype. Leppert and coworkers proposed a model of leiomyoma development based on an abnormal response to tissue repair, resulting in disordered healing, myofibroblast transformation, myofibroblast failure to undergo apoptosis and formation of an altered ECM (Leppert et al. 2006). In order to shed light onto the origin of these uterine myofibroblasts, we investigated whether they derive from a population of resident stromal fibroblasts identified by CD34positivity. Lindenmayer and Miettinen reported the presence of CD34-positive fibroblastic cells in the myometrium (Lindenmayer and Miettinen 1995) and speculated that they serve as myofibroblast precursors. CD34, a 110-KDa transmembrane cell surface glycoprotein, was first described in hematopoietic progenitor cells (Strauss et al. 1984) and later in other cell types (Nielsen and McNagny 2008). CD34 identifies a subset of fibroblasts located in both perivascular and stromal positions in the connective tissue of multiple anatomic sites (Diaz-Flores et al. 2014). Furthermore, CD34-positive fibroblastic cells behave as mesenchymal stem cell progenitors and play an important role in wound healing, tissue repair, scarring and tumor stroma formation (Barth et al. 2002; Barth and Westhoff 2007). When activated, these cells lose CD34 expression and may acquire α-sma expression giving rise to myofibroblasts. Depending on location and cell characteristics, this cell type has received several different names, including CD34+ fibroblasts, CD34+ fibrocytes (including tissue resident and circulating CD34+ fibrocytes), CD34+ spindle cells (or CD34+ spindle-shaped cells), CD34+ reticular network cells and other names leading to confusion (for reviews, see Diaz-Flores et al. 2014; Erdag et al. 2007). Here, we used an antibody directed against CD34 to identify a population of stromal resident cells, which are characterized by an ovoid and creased nucleus and a thin layer of bipolar or multipolar cytoplasmic processes (Diaz-Flores et al. 2014). To understand how inflammation and fibrogenesis contribute to leiomyoma pathogenesis, we first investigated the presence and location of inflammatory cells inside leiomyomas, in the immediate surrounding myometrium and in normal myometrium at least 1.5 cm from the border of leiomyomas (defined as Bdistant^). Next, we studied the localization of the ECM, the expression of α-sma and desmin and the immunohistochemical profile of the mesenchymal cells that were present inside the leiomyoma and in the surrounding myometrium. Finally, based on the previous findings that activin A plays a pro-fibrotic role and is highly expressed in uterine leiomyomas (Ciarmela et al. 2011a, b; Islam et al. 2014), we evaluated whether and which inflammatory molecules are able to

Cell Tissue Res (2016) 364:415–427

stimulate the expression of this profibrotic cytokine in leiomyoma and myometrial cells in vitro.

Materials and methods Patients and samples The subjects included in this study were 24 women submitting to total hysterectomy for the presence of numerous leiomyomas of variable sizes, diagnosed by ultrasonography before surgery. The study was approved by the local Ethics Committee and informed consent was obtained. After hysterectomy, uterus tissue was immersion-fixed for 24–48 h in buffered formalin and sent to a pathology laboratory for pathological reporting. Scheduled sampling was performed. In the present study, samples of leiomyomas, the surrounding myometrium and, when it was available, the myometrium distant from leiomyomas were used. The myometrial tissue distant from the leiomyoma was from the corpus or fundus, at least 1.5 cm from any leiomyoma present in the uterus and the evaluated areas were in the central part of the myometrium, distant from the endometrium and the external serous surface. A single leiomyoma was selected for the study from all patients. Leiomyomas with degenerative aspects such as hemorrhage, necrosis, edema and/or calcifications were excluded. Cases featuring inflammatory pathologies, such as acute and chronic endometritis or peritonitis, as well as samples taken from women who had received exogenous hormones in the previous 3 months, were also excluded. Leiomyoma tissues were defined using well-established histopathologic criteria (Tavassoli 2003). The study included three different types of leiomyomas: (1) usual (n=15), (2) cellular (n=7) and (3) lipoleiomyoma (n=2). Leiomyomas were of variable sizes ranging from 2 to 12 cm in diameter. All collected samples were paraffin-embedded for subsequent histopathological and immunohistochemical investigations. Antibodies To characterize the immunohistochemical profile of the inflammatory infiltrate, we used the following monoclonal antibodies purchased from Novocastra™ (Newcastle upon Tyne, UK): CD45 or Leukocyte Common Antigen, LCA (clone X16/99, ready to use), which labels the majority of inflammatory cells; CD68 (KP1, diluted 1:1500), against tissue macrophages; and Mast Cell Tryptase, MCT (10D11, ready to use), which distinguishes the granules of mast cells. To study the immunoprofile of the leiomyoma cells, we used α-sma (Novocastra™, αsm-1, diluted 1:50), a component of the cellular cytoskeleton that is present in large amounts in all the human cells with contractile properties apart from voluntary muscles and desmin (Novocastra™, DE-R-11, ready

417

to use), a cytoskeletal protein characteristic of smooth and skeletal muscle myocytes. We also used a monoclonal antibody against CD34 (Novocastra™, QBEnd/10, ready to use) to highlight CD34+ stromal fibroblasts. TNF-α (clone ab9739, diluted 1:400) for immunohistochemical study was purchased from Abcam (Cambridge, UK). Activin A (rabbit, anti-βA, 81–113)-NH2 anti-serum diluted 1:2500 was kindly donated by Dr Joan Vaughan and Dr Peter Gray (Salk Institute for Peptide Biology, La Jolla, CA, USA) and were used as previously described (Bloise et al. 2010; Ferreira et al. 2008). Immunohistochemistry The Bond™ Polymer Refine Detection Kit (DS9800; Leica Biosystems, Novocastra™) was applied using the mouse primary monoclonal antibody for 30 min at the dilutions mentioned above. The procedure was performed according to the manufacturer’s instructions. After counterstaining with hematoxylin, the sections were dehydrated through 80 %, 95 %, absolute alcohol and xylene. Serum instead of the primary antibody was used as negative control. Appropriate positive controls were used: normal human tonsil tissue for CD68, CD45 and MCT; for α-sma, normal human gastrointestinal tissue; for desmin, normal human muscle; and for TNFalpha and activin A, human placental tissue (Petraglia et al. 1992). Conveniently, α-sma, desmin and CD34 have internal positive controls in all examined samples, i.e., the vessel wall. Tissue sections were mounted with resinous mounting medium. CD68-, CD45- and MCT-positive cells were counted in five different images (×200 magnification) carried out by AxioVision Rel. 4.8. Results are presented as cell numbers present in 0.17 mm2. For accuracy, analyses were performed by two individuals in a blinded fashion. Masson’s trichrome Masson’s trichrome stain was used to study consistency, disposition and amount of ECM (Luna 1968). The formalin-fixed and paraffin-embedded uterine leiomyoma and myometrial tissue sections were deparaffinised and rehydrated. The sections were washed in distilled water and post-fixed with Bouin’s solution overnight. The sections were then rinsed in running tap water for 10 min and stained in Weigert’s iron hematoxylin working solution for 10 min. They were rinsed in running warm tap water for 10 min and then washed in distilled water. Sections were counter-stained in Biebrich’s scarlet-acid fuchsine solution for 30 s and washed with distilled water. Tissue sections were then differentiated in phosphomolybdic–phosphotungstic solution for 15 min. They were transferred directly (without rinsing) to light green solution and stained for 2–6 min, rinsed in distilled water and differentiated in 1 % acetic acid solution for 1 min. Then, sections were washed in distilled water and dehydrated very

418

quickly through 95 %, absolute alcohol and xylene and mounted with resinous mounting medium. All the above reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA). Quantitative morphometric analysis was performed as previously reported (Haller et al. 2012). Collagen expression was quantified by measuring the area of five fields of Masson's trichrome staining using Adobe Photoshop 7.0 and Image J 1.49n (National Institutes of Health, USA http://imagej.nih. gov/ij). Random images were captured from collagenstained slides using a Nikon Y-THS microscope (Japan). The area of collagen was calculated from the images by performing sequential commands in Photoshop (open image>select>color range>fix fuzziness less then 100>select blue-green stained color from image>ok). For white areas (vessels lumen and broken parts of the tissue), fuzziness was set as less than 50. Selected parts were copied and pasted into a new page and saved in Tiff format. The Tiff images were opened in Image J and the scale set as distance in pixels: 317, known distance: 200, pixel aspect ratio: 1.0 and unit of length: μm. Tiff images were converted into a RGB stack and then the threshold adjusted. Area, area fraction and limit to threshold were selected as parameters. Then, analyze>measure, the percentage of the area was visible in a new window and saved as an Excel file. The percentage of white area was removed from the total area (100) and the percentage of collagen-occupied area was recalculated. The final value was obtained by averaging the five fields. Sirius red Sirus Red Puchtler from Bio-Optica (Milan, Italy) was used to stain for collagen and the procedure was performed according to the manufacturer’s instructions. Type I collagen fibers are orange to red, whereas the thinner type III collagen fibers appear yellow to green.

Cell Tissue Res (2016) 364:415–427

cell lines were generated from primary cultures by immortalization using HPV-16 as previously described (Malik et al. 2008). Real-time PCR Cells were lysed using TRIzol® reagent (Life Technologies Invitrogen, Carlsbad, CA, USA) and stored at −80 °C. Total RNA was isolated using chloroform according to the manufacturer’s instructions. The ReliaPrep™ RNA Cell Miniprep System was used to purify and concentrate RNA (Promega, Madison, WI, USA). cDNA was generated from 1 μg total RNA by reverse transcription using a high-capacity cDNA RT kit (Applied Biosystems). Real-time PCR was performed using 50 ng cDNA with the following thermal cycle protocol: initial denaturation at 95 °C for 20 s, followed by 40 cycles of 95 °C for 1 s and 60 °C extension for 20 s. The following TaqMan gene expression assays (Applied Biosystems) were used: activin A subunit (Hs00170103_m1), HPRT (Ciarmela et al. 2014), (Hs99999909_m1) and β Actin (Hs99999903_m1), the latter two being used as housekeeping genes. The blank for each reaction, consisting in amplifications performed in the absence of the RT enzyme, was performed. Statistical analysis Statistical analyses were performed using GraphPad Prism v. 6.01 (GraphPad, San Diego, CA). For paired comparison among a leiomyoma and either corresponding myometrium close and distant, data were analyzed using Friedman’s test, followed by a post-hoc Dunn’s multiple comparison test. To compare the usual versus cellular type of leiomyoma and to compare the effect of the inflammatory molecules in the in vitro experiments, the data were analyzed using a Kruskal– Wallis test, followed by a post-hoc Dunn’s multiple comparison test. Differences were considered significant when P