Successful grafting of tissue-engineered fetal skin.

PURPOSE: Fetal repair of spina bifida results in improved outcomes and has therefore become a standard clinical procedure in some highly specialized centers. However, optimization of the procedure technique and timing is needed. Both might be achieved by facilitating the procedure using laboratory-grown fetal skin substitutes. The aim of this study was therefore to test in vivo the suitability of such a fetal skin substitute for an in utero application. METHODS: Collagen-based hydrogels containing fetal ovine fibroblasts were seeded with fetal ovine keratinocytes and transplanted on immuno-incompetent nu/nu rats. After 3 weeks, grafts were harvested and analyzed histologically and by immunohistochemistry. RESULTS: Laboratory-grown fetal ovine dermo-epidermal skin substitutes showed successful engraftment at 3 weeks. Histologically, grafts revealed a neo-dermis populated by fibroblasts and with ingrowth of vessels, and an epidermis with an adult-like, mature appearance depicting clearly basal, spinous, granular, and a corneal layer. Immunostaining confirmed a physiologically organized epidermis. CONCLUSION: Fetal dermo-epidermal skin substitutes of ovine origin can successfully be grafted in vivo. In a next step, we will have to test whether favorable results can also be obtained when grafts are used in utero. If so, then human fetal spina bifida repair using laboratory-grown autologous fetal skin for defect closure may be envisaged.

Myelinated and unmyelinated nerve fibers reinnervate tissue-engineered dermo-epidermal human skin analogs in an in vivo model.

PURPOSE: The clinical application of autologous tissue-engineered skin analogs is an important strategy to cover large skin defects. Investigating biological dynamics, such as reinnervation after transplantation, is essential to improve the quality of such skin analogs. Previously, we have examined that our skin substitutes are reinnervated by host peripheral nerve fibers as early as 8 weeks after transplantation. Here, we wanted to investigate the presence and possible differences regarding myelinated and unmyelinated host nerve fibers 15 weeks after the transplantation of light and dark human tissue-engineered skin analogs. METHODS: Human epidermal keratinocytes, melanocytes, and dermal fibroblasts were isolated from human light and dark skin biopsies. Keratinocytes and melanocytes were seeded on fibroblast-containing collagen type I hydrogels after expansion in culture. After additional culturing, the tissue-engineered dermo-epidermal skin analogs were transplanted onto full-thickness skin wounds created on the back of immuno-incompetent rats. Skin substitutes were excised and analyzed 15 weeks after transplantation. Histological sections were examined with regard to the ingrowth pattern of myelinated and unmyelinated nerve fibers into the skin analogs using markers, such as Substance P, NF200, and S100-Beta. RESULTS: We found myelinated and unmyelinated peripheral host nerve fibers 15 weeks after transplantation in the dermal part of our human skin substitutes. In particular, we identified large-diameter-myelinated Aß- and Aδ-fibers, and small-diameter C-fibers. Furthermore, we observed myelinated nerves in close proximity to CD31-positive blood capillaries. In the long run, both types of ingrown host fibers showed an identical pattern in both light and dark skin analogs. CONCLUSION: Our data suggest that myelinated and unmyelinated peripheral nerves reinnervate human skin substitutes in a long-term in vivo transplantation assay. Our tissue-engineered skin analogs attract A- and C-fibers to supply both light and dark skin analogs. Potentially, this process restores skin sensitivity and has, therefore, a significant relevance with regard to future application of autologous pigmented dermo-epidermal skin substitutes onto patients.

Experimental tissue engineering of fetal skin.

PURPOSE: In some human fetuses undergoing prenatal spina bifida repair, the skin defect is too large for primary closure. The aim of this study was to engineer an autologous fetal skin analogue suitable for in utero skin reconstruction during spina bifida repair. METHODS: Keratinocytes (KC) and fibroblasts (FB) isolated from skin biopsies of 90-day-old sheep fetuses were cultured. Thereafter, plastically compressed collagen hydrogels and fibrin gels containing FB were prepared. KC were seeded onto these dermal constructs and allowed to proliferate using different culture media. Constructs were analyzed histologically and by immunohistochemistry and compared to normal ovine fetal skin. RESULTS: Development of a stratified epidermis covering the entire surface of the collagen gel was observed. The number of KC layers and degree of organization was dependent on the cell culture media used. The collagen hydrogels exhibited a strong tendency to shrink after eight to ten days of culture in vitro. On fibrin gels, we did not observe the formation of a physiologically organized epidermis. CONCLUSION: Collagen-gel-based ovine fetal cell-derived skin analogues with near normal anatomy can be engineered in vitro and may be suitable for autologous fetal transplantation.

New mammary epithelial and fibroblastic cell clones in coculture form structures competent to differentiate functionally.

We have established and characterized a spontaneously immortalized, nontumorigenic mouse mammary cell line, designated IM-2. IM-2 cells synthesize large amounts of the milk protein beta-casein upon addition of lactogenic hormones. The induction of beta-casein occurs rapidly and does not require any exogenous extracellular matrix components. The IM-2 cell line is morphologically heterogeneous and could be separated into cell clones with epithelial and fibroblastic characteristics. In monoculture, none of the epithelial clones could be induced to synthesize caseins. Coculture of epithelial and fibroblastic clones, however, rendered the epithelial cells competent to differentiate functionally; the addition of lactogenic hormones to these cocultures resulted in the synthesis of beta-casein in amounts comparable to that seen with the original IM-2 line. Using this unique cell system, we have investigated the interrelationships between different steps in differentiation leading to hormone-induced casein production. Independent of hormones, epithelial-fibroblastic cell contacts led to the formation of characteristic structures showing the deposition of laminin. We found that the epithelial cells located in these structures also exhibited significantly increased levels of cytokeratin intermediate filament polypeptides. Double immunofluorescence revealed that the cells inducible by hormones to synthesize casein, colocalized exactly with the areas of laminin deposition and with the cells showing greatly intensified cytokeratin expression. These results suggest that hormone-independent differentiation events take place in response to intercellular epithelial-mesenchymal contacts. These events in turn bring about a state of competence for functional differentiation after lactogenic hormonal stimulation.

The estrogen-dependent c-JunER protein causes a reversible loss of mammary epithelial cell polarity involving a destabilization of adherens junctions.

Members of the epidermal growth factor (EGF) receptor family are known to be specifically involved in mammary carcinogenesis. As a nuclear target of activated receptors, we examined c-Jun in mammary epithelial cells. For this, we used a c-JunER fusion protein which was tightly controlled by estrogen. Activation of the JunER by hormone resulted in the transcriptional regulation of a variety of AP-1 target genes. Hormone-activated JunER induced the loss of epithelial polarity, a disruption of intercellular junctions and normal barrier function and the formation of irregular multilayers. These changes were completely reversible upon hormone withdrawal. Loss of epithelial polarity involved redistribution of both apical and basolateral proteins to the entire plasma membrane. The redistribution of E-cadherin and beta-catenin was accompanied by a destabilization of complexes formed between these two proteins, leading to an enrichment of beta-catenin in the detergent-soluble fraction. Uninduced cells were able to form three-dimensional tubular structures in collagen I gels which were disrupted upon JunER activation, leading to irregular cell aggregates. The JunER-induced disruption of tubular structures was dependent on active signaling by growth factors. Moreover, the effects of JunER could be mimicked in normal cells by the addition of acidic fibroblast growth factor (aFGF). These data suggest that a possible function of c-Jun in epithelial cells is to modulate epithelial polarity and regulate tissue organization, processes which may be equally important for both normal breast development and as initiating steps in carcinogenesis.

A simplified fabrication technique for cellularized high-collagen dermal equivalents.

Human autologous bioengineered skin has been successfully developed and used to treat skin injuries in a growing number of cases. In current clinical studies, the biomaterial used is fabricated via plastic compression of collagen hydrogel to increase the density and stability of the tissue. To further facilitate clinical adoption of bioengineered skin, the fabrication technique needs to be improved in terms of standardization and automation. Here, we present a one-step mixing technique using highly concentrated collagen and human fibroblasts to simplify fabrication of stable dermal equivalents. As controls, we prepared cellularized dermal equivalents with three varying collagen compositions. We found that the dermal equivalents produced using the simplified mixing technique were stable and pliable, showed viable fibroblast distribution throughout the tissue, and were comparable to highly concentrated manually produced collagen gels. Because no subsequent plastic compression of collagen is required in the simplified mixing technique, the fabrication steps and production time for dermal equivalents are consistently reduced. The present study provides a basis for further investigations to optimize the technique, which has significant promise in enabling efficient clinical production of bioengineered skin in the future.

Impact of carbon dioxide versus air pneumoperitoneum on peritoneal cell migration and cell fate.

BACKGROUND: Postoperative systemic immune function is suppressed after open abdominal surgery, as compared with that after minimally invasive abdominal surgery. As a first line of defense, peritoneal macrophages (PMo) and polymorphonuclear neutrophil granulocytes (PMNs) are of primary importance in protecting the body from microorganisms. Previous studies have shown changes in these cell populations over time after open versus laparoscopic surgery. This study aimed to investigate the dynamics of cell recruitment and clearance of peritoneal cells. METHODS: Female NMRI mice (33 +/- 2 g) were randomly assigned to carbon dioxide (CO2) or air insufflation. Intravasal cells with phagocytic capabilities were selectively stained by intravenous injection of the fluorescent dye PKH26 24 h before surgery. Gas was insufflated into the peritoneal cavity through a catheter, and the pneumoperitoneum was maintained for 30 min. Peritoneal lavage was performed 1, 3, 8, or 24 h after surgery. Apoptotic cells were assessed by flow cytometry using a general caspase substrate. RESULTS: The total peritoneal cell count did not differ between groups. The PKH26-positive PMo level was significantly increased after CO2, as compared with air, at 1 h and 24 h. The ratio of apoptotic PMo did not differ between the groups. In the peritoneal lavage, polymorphonuclear leukocytes (PMNs) were tripled in the air group, as compared with the CO2 group, whereas the ratio of apoptotic PMNs was significantly decreased. There was a higher fraction of PKH26-positive PMNs after air exposure, as compared with that after CO2. CONCLUSIONS: Air exposure triggered a higher transmigration rate of PMNs from the blood compartment into the peritoneal cavity and decreased PMN apoptosis, as compared with CO2. The lower proportion of PKH26-positive peritoneal macrophages in the air group might have been attributable to a higher inflammatory stimulation than in the CO2 group, leading to increased emigration of PMo to draining lymph nodes. All the findings underscore a complex cell-specific regulation of cell recruitment and clearance in the peritoneal compartment.

T1, an immunoglobulin superfamily member, is expressed in H-ras-dependent epithelial tumours of mammary cells.

T1 is a glycosylated protein in the carcinoembryonic antigen (CEA) family of tumour marker molecules. It was originally identified by virtue of its transient induction after the expression of p21H-ras in NIH3T3 fibroblasts. Here we show that the T1 gene is activated in mammary adenocarcinomas of transgenic mice harbouring an H-ras transgene under the control of the mammary-specific whey acidic protein (WAP) promoter. By contrast, T1 mRNA was not, or only faintly, detectable in mammary carcinomas of transgenic mice bearing a WAP-myc transgene. Thus, T1 overexpression does not appear to be a general tumour-specific phenomenon. A dependence of T1 gene expression on the action of p21H-ras is suggested by the observation of T1 mRNA in nude mouse tumours generated from H-ras-transformed cultured mammary epithelial cells. Interestingly, activation of the T1 gene is also found during the maturation of the mammary gland (3-4 weeks after birth), whereas it is absent during its terminal differentiation in pregnancy and lactation. This expression pattern suggests a role for the secreted T1 glycoprotein in the phase of epithelial proliferation of the mammary gland. It appears that p21H-ras-induced transformation of mammary epithelial cells mimics the situation occurring in puberty. In both developmental stages the T1 glycoprotein might affect cell interactions of the proliferating epithelial cells with the surrounding stroma. It might thus promote ductal outgrowth in gland maturation as well as invasive growth of p21H-ras-transformed mammary epithelial cells.

Progression of carcinoma cells is associated with alterations in chromatin structure and factor binding at the E-cadherin promoter in vivo.

E-cadherin has been identified as a tumor (invasion) suppressor gene, which is mutated in 50% of diffuse-type human gastric carcinomas. In other carcinomas, the expression of E-cadherin is down-regulated in the poorly differentiated cells such as from breast, bladder, lung and colon. We have here examined the in vivo properties of the genomic E-cadherin promoter in well and poorly differentiated carcinoma cell lines in order to gain insights into the mechanisms of E-cadherin down-regulation in tumors. In vivo footprinting analysis revealed that positive regulatory elements of the E-cadherin promoter (a GC-rich region, the CCAAT-box and a palindromic element) are specifically bound by transcription factors in E-cadherin-expressing but not in non-expressing cells. The tested cell systems include more than a dozen carcinomas cell lines as well as mammary epithelial cells where E-cadherin expression can be switched off by activation of a Fos-estrogen receptor fusion protein and rhabdomyosarcoma cells where E-cadherin expression was induced by transfection with E1A. Mapping of DNase I hypersensitive sites showed that the chromatin structure in the promoter region is loosened in expressing but condensed in non-expressing cells. Furthermore, the endogenous E-cadherin promoter is specifically methylated at CpG sites in the undifferentiated cells. We also show that the in vivo properties of the promoter in E-caherin-negative carcinoma cells are similar as in mesenchymal cells, i.e. fibroblasts or sarcoma cells. These data suggest that silencing of the E-cadherin promoter during epithelialmesenchymal transition and tumor progression is due to a loss of factor binding in vivo and to chromatin rearrangement in the regulatory region.

Fas-dependent tissue turnover is implicated in tumor cell clearance.

The apoptosis-inducing Fas receptor has been shown to be down-regulated in various types of tumors, while its ligand (FasL) appears to be frequently up-regulated. Here we provide evidence that there is a strong selective pressure in vivo against Fas-expressing, tumorigenic NIH3T3 cells, favoring survival, proliferation and eventually tumor formation by Fas-negative cells. Importantly, re-expression of Fas in these cells results in either the complete abolishment of tumor development, or in a significant extenuation of the latency period of tumor outgrowth. In addition, we found that environmental conditions which prevail during tumorigenesis, such as limiting amounts of survival factors and the lack of cell adhesion, are markedly sensitizing tumor cells to Fas-mediated suicide. Our data suggest that in addition to T cell-mediated immune responses, mechanisms of Fas-dependent tissue turnover are also centrally implicated in tumor cell clearance.

The transformation of primary and established mouse mammary epithelial cells by p21-ras is concentration dependent.

We constructed a retroviral vector (pZSR) which is capable of simultaneously expressing the neomycin resistance gene and the viral ras oncogene. Primary mammary gland epithelial cells were prepared from mid-pregnant mice and infected with this virus. Cell lines with epithelial cell characteristics could be established with a low frequency. High expression of p21 v-ras was observed in these cells. They are tumorigenic and form soft agar colonies dependent on the presence of EGF and insulin in the growth medium but progress to hormone independent growth at higher passage numbers. A cloned cell line of non-tumorigenic, established mammary epithelial cells (NOG8) was also infected with the v-ras expressing virus. Individual cell clones expressing increasing amounts of p21 v-ras were selected. The level of p21 v-ras expression directly influences the morphology of the epithelial cells in culture, determines their cloning efficiency in soft agar and their tumorigenicity.

Early peritoneal macrophage function after laparoscopic surgery compared with laparotomy in a mouse mode.

BACKGROUND: The authors previously demonstrated postoperative preservation of the immune function measured by delayed-type skin reaction and tumor growth after laparoscopic surgery, as compared with laparotomy. For further elucidation of the origin of the demonstrated immune preservation, peritoneal macrophage (PMo) function was investigated 1 h after different surgical procedures. METHODS: Female NMRI mice were divided into five groups: anesthesia only, abdominal skin incision, laparotomy, peritoneal carbon dioxide (CO2) insufflation, and peritoneal air insufflation. Escherichia Coli phagocytosis, tumor necrosis factor-alpha (TNF-alpha), transforming growth factor-beta1 (TGF-beta1), and interleukin-10 (IL-10) release of isolated PMo were investigated. RESULTS: All invasive interventions reduced the PMo phagocytosis by factors of approximately 2 to 4.7, as compared with the sham control group. Spontaneous ex vivo TNF-alpha release was significantly increased whenever the abdominal cavity was exposed to ambient air. The macrophage's ability to release TNF-alpha after E. coli exposure was diminished in the abdominal air exposure groups, as compared with the CO2 insufflation group. CONCLUSION: Reduced phagocytosis 1 h after surgical interventions suggests a contribution of PMo to the altered immune function. When exposed to CO2, PMo show a decreased basal TNF-alpha release. However, PMo also show an increased TNF-alpha release after a second immune stimulation (E. coli), suggesting a greater competency of interaction in an immune defense reaction after CO2 exposure.

v-myc alters the response of a cloned mouse mammary epithelial cell line to lactogenic hormones.

Several oncogenes have now been implicated in mammary carcinogenesis. We investigated the phenotypic effects of expressing three representative oncogenes in mammary epithelial cells. v-myc (coding for a nuclear protein), v-Ha-ras (a G-protein homologue) and v-fgr (a tyrosine kinase) genes were introduced into the nontumorigenic clone 14 of the mouse mammary epithelial cell line COMMA-1D. Their effects upon growth and differentiation were determined. Anchorage-independent growth was induced by all three oncogenes with low efficiency. v-Ha-ras and v-fgr induced tumorigenicity in nude mice. The effect of oncogenes upon parameters unique to mammary epithelial cells in vitro was assayed. Both v-myc and v-fgr abolished the ability of clone 14 to grow as three-dimensional branching structures in hydrated collagen gel. v-fgr completely and v-myc partially inhibited the expression of the epithelium specific cytokeratins. Clone 14 can be induced to produce the beta-casein milk protein by the combination of the lactogenic hormones, dexamethasone, insulin, and PRL. Introduction of v-myc into clone 14 cells resulted in an estimated 50-fold increased induction of beta-casein protein and at least a 60-fold increase in beta-casein mRNA. The number of cells stained with anti-beta casein antibodies also showed a 10-fold increase after v-myc introduction. This still required the synergistic action of all three lactogenic hormones. Thus v-myc can alter the normal response of mammary epithelial cells to lactogenic hormones.

Oncoprotein fos activation in epithelial-cells induces an epitheliomesenchymal conversion and changes the receptor encoded by the fgfr-2 messenger-RNA from k-sam to bek.

Activation of c-Fos, by using an inducible c-Fos estrogen receptor fusion protein, triggers the epitheliofibroblastoid cell conversion of mouse mammary epithelial cells. We show that this change in phenotype is accompanied by a definitive switch of the fibroblast growth factor receptor 2 from K-SAM to BEK. This splicing switch occurs a few hours after estrogen stimulation. Our data suggest that Fos proteins could be important in modulating the FGFR-2 splicing choice. Moreover, these observations reinforce previous evidence that the BEK/K-SAM choice is strictly tissue-specific: the K-SAM exon is expressed exclusively in epithelial cells, the BEK exon in cells of the fibroblastic type.

Novel treatment for massive lower extremity avulsion injuries in children: slow, but effective with good cosmesis.

BACKGROUND: Extended avulsion injuries are associated with significant loss of skin and subcutaneous fat, leaving the reconstructive surgeon with the challenge of substituting all tissues lost in the best possible way. We wanted to test whether the combined use of a Vacuum Assisted Closure system (VAC) and Integra Dermal Regeneration Template (IDRT) matched the required treatment profile encompassing initial control of infection, remodeling of body contours, and reconstruction of near normal skin. MATERIALS AND METHODS: 4 children with massive lower extremity avulsion injuries were treated with early necrosectomy, VAC application for 3-5 weeks for wound cleansing and wound bed conditioning, subsequent implantation of IDRT, and finally autologous split thickness skin grafting (STSG) for definitive wound closure. Thereafter, a standard rehabilitation program was used. The key parameters of VAC and IDRT application, take rates of IDRT and STSG, complications, length of stay, and final outcome were recorded. RESULTS: In all patients, early removal of necrosis and infection control was successfully achieved. Continuous VAC application fostered the formation of a several millimeters thick new tissue layer partly compensating for the lost hypodermis. IDRT implantation and subsequent STSG yielded take rates of nearly 100% and both functionally and cosmetically excellent long-term results. There were no major complications. CONCLUSION: The combination of VAC and IDRT in children with massive leg avulsion injuries is feasible, safe, and delivers high-quality long-term outcomes that appear to justify the multiple operative procedures, the long hospitalization times, and the comparatively high costs entailed.

Osmotic expanders in children: no filling--no control--no problem?

BACKGROUND: Self-filling, hydrogel-based osmotic tissue expanders have been successfully used for several years, mainly in adult patients. We wanted to test this novel device in pediatric plastic and reconstructive surgery. MATERIAL AND METHODS: Between November 2004 and September 2009, we implanted 53 osmotic tissue expanders following standard surgical principles in a total of 30 children and adolescents with burn scars, congenital nevi, alopecia, or foot deformities. RESULTS: All expanders reached their predicted volume within 6 weeks and 51 (96.2%) produced a sufficient amount of additional skin for the intended coverage of the defect. A serious infection precluding the planned reconstructive procedure occurred with 2 expanders (3.8%). Minor complications occurred at 6 implantation sites (11.4%), and consisted of small necrotic areas and perforations (n = 3) and minor infections (n = 3). These problems could be controlled and did not interfere with the subsequent plasty. The final results recorded at the last follow-up (mean: 21 months, range: 9-48 months) were rated as excellent in 25, good in 19, moderate in 6, and poor in 1 patient. CONCLUSION: This is apparently the largest pediatric series in which self-filling expanders have been used. The data obtained indicates that self-filling expanders can be safely and effectively used for various plastic, reconstructive and orthopedic procedures in children and adolescents. The fact that numerous painful and distressing filling sessions are obviated with these expanders is particularly beneficial for those children too young to understand and cooperate. Moreover, this approach minimizes the risk of infection and lowers costs.

Functional characterization of cell hybrids generated by induced fusion of primary porcine mesenchymal stem cells with an immortal murine cell line.

Bone marrow mesenchymal stem cells (MSC) integrate into various organs and contribute to the regeneration of diverse tissues. However, the mechanistic basis of the plasticity of MSC is not fully understood. The change of cell fate has been suggested to occur through cell fusion. We have generated hybrid cell lines by polyethylene-glycol-mediated cell fusion of primary porcine MSC with the immortal murine fibroblast cell line F7, a derivative of the GM05267 cell line. The hybrid cell lines display fibroblastic morphology and proliferate like immortal cells. They contain tetraploid to hexaploid porcine chromosomes accompanied by hypo-diploid murine chromosomes. Interestingly, many hybrid cell lines also express high levels of tissue-nonspecific alkaline phosphatase, which is considered to be a marker of undifferentiated embryonic stem cells. All tested hybrid cell lines retain osteogenic differentiation, a few of them also retain adipogenic potential, but none retain chondrogenic differentiation. Conditioned media from hybrid cells enhance the proliferation of both early-passage and late-passage porcine MSC, indicating that the hybrid cells secrete diffusible growth stimulatory factors. Murine F7 cells thus have the unique property of generating immortal cell hybrids containing unusually high numbers of chromosomes derived from normal cells. These hybrid cells can be employed in various studies to improve our understanding of regenerative biology. This is the first report, to our knowledge, describing the generation of experimentally induced cell hybrids by using normal primary MSC.

Oncogenes and epithelial cell transformation.

More than 80% of the tumors occurring in man are carcinomas. Despite the prevalence of these epithelial tumors most in vitro studies of oncogenesis have employed mesenchymal rather than epithelial cells. As a result, a detailed understanding of the molecular mechanisms of carcinoma formation and an exact definition of the malignant transformed epithelial phenotype are lacking. Taking into account that the genesis of a carcinoma involves alterations in multiple genes, giving rise to complex phenotypic changes, the question arises whether each single, mutational step can be correlated with distinct alterations of the epithelial phenotype. To approach this question, oncogenes that are both critically positioned in growth factor receptor-driven signaling pathways and relevant in carcinoma development, such as ras, src, myc and fos, have been expressed in epithelial cells. The results indicate that epithelial cells transformed by oncogenes in vitro alter their gene expression programs, thereby losing certain features of cell polarity and cell adhesion. However, as with true tumor cells, these changes occur in distinct combinations, to different degrees and are influenced by the local environment of the cells.

Two-dimensional gel mapping of small GTPases reveals transformation-specific changes during oncogenesis.

Epithelial cells transformed by oncogenes in vitro change their gene expression program, thereby losing features of cell polarity and cell adhesion. Using ras-transformed mammary epithelial cells, we have investigated the expression of other small GTP-binding proteins by high-resolution two-dimensional gel electrophoresis and direct GTP ligand binding after renaturing transfer onto nitrocellulose. Ras-transformed cells lost the expression of one epithelial-specific GTP-binding protein (21-22 kDa, pI 4.5-4.8) and instead expressed a fibroblast GTP-binding protein (21-22 kDa, pI 4.8-5.0).