Uncoupling Keratocyte Loss of Corneal Crystallin from Markers of Fibrotic Repair

Abstract

purpose. Corneal crystallins are lost from resident cells of the corneal stroma during wound repair, and this is associated with a loss of cell transparency. The goal of this study was to identify factors inducing loss of the corneal crystallin transketolase (TKT).

methods. A cell culture model of freshly isolated rabbit corneal keratocytes was used. Fibrotic markers included cell proliferation, adoption of a “fibroblastic” spindle-shaped morphology associated with cytoskeletal rearrangement, loss of TKT, and expression of α-smooth muscle actin (α-sm actin), a marker for the myofibroblast.

results. When freshly isolated keratocytes were cultured in the continuous presence of 10% calf serum, the high level of intracellular TKT protein was reduced dramatically within 24 to 48 hours. In contrast, TKT protein was retained in cells maintained in the absence of serum. When cells were prevented from proliferating by exposure to serum for <24 hours or by continuously exposing to serum at a contact-inhibiting plating density, TKT loss was inhibited. TKT loss was induced by treatment of serum-free cultures with the serum cytokines platelet-derived growth factor or basic fibroblast growth factor, both of which also stimulated keratocyte proliferation, although not other changes associated with fibrosis. However, TKT loss was not induced by treatment of serum-free cultures with a third serum cytokine, transforming growth factor- (TGF)-β, even though TGF-β stimulated cell proliferation at low doses and induced the fibroblastic spindle-shape and express α-sm actin at high doses.

conclusions. TKT loss in corneal keratocytes can be induced by PDGF or bFGF and this loss can be uncoupled from other fibrotic markers. Targeting these cytokines or the signaling pathways that they activate could enable retention of corneal crystallin in stromal cells during repair, a more regenerative outcome. The result would be enhanced clarity of the cornea.

After injury to connective tissues such as the dermis of the skin or the corneal stroma, cells at the wound periphery become metabolically activated, enter the cell cycle, migrate into the damaged area, and transition to a repair cell type that shares common phenotypic characteristics across organs. ^{1 2} These cells proliferate and deposit a disorganized, fibrotic extracellular matrix (ECM) that seals off the damaged area, much like a spot weld. However, repair is imperfect, because repair tissue can never reproduce the functions of the original tissue that it replaces. Understanding the molecular pathways controlling fibrotic repair may enable the design of strategies to improve the regenerative character of the repair process. 

In the cornea, repair cells are derived from the resident keratocytes of the corneal stroma. ^{3 4 5} The phenotypic transition from keratocyte to repair cell in corneal wounds is similar to the change that occurs when keratocytes are isolated from the normal stroma, placed in cell culture, and exposed to serum. ^{2 6 7 8 9} The transition is characterized by (1) a change in cell shape from dendritic to spindle-shaped associated with reorganization of the actin cytoskeleton to form stress fibers, ⁷ (2) induction of α5-integrin expression and its heterodimerization with β1-integrin to form the fibronectin receptor, ^{7 10} (3) deposition of a repair-type extracellular matrix (ECM) through induced expression of new molecule such as fibronectin and SPARC, ^{7 11} and (4) competence to activate an IL-1α autocrine feedback loop essential for the control of collagenase expression and subsequent repair tissue remodeling. ^{12 13 14} Cells in culture with these characteristics are called fibroblasts and, as just stated, they are similar to the fibroblastic cells present in corneal repair tissue in vivo. In cornea, the repair transition is also associated with loss of specific keratocyte markers, including secreted keratan sulfate proteoglycans (KSPGs) such as prostaglandin D synthase, ^{8 15} and the intracellular corneal crystallins, which in the rabbit are the metabolic enzymes transketolase (TKT) and aldehyde dehydrogenase class 1A1 (ALDH1A1). ¹⁶ Loss of the latter proteins is particularly significant, as they are thought to help confer transparency to corneal stromal cells. Under the influence of cytokines of the TGF-β family, a subset of repair fibroblasts may also begin to express a new gene product, α-smooth muscle actin (α-sm actin). ^{10 17} Cells expressing α-sm actin are often called myofibroblasts. Expression of α-sm actin is often also associated with a larger cell size and increased organization of filamentous actin into stress fibers. These molecular characteristics contribute to cell contractility, facilitating wound contraction. ¹⁸  

Cell culture can be used to screen for factors with the capacity to convert keratocytes to repair cells. Keratocytes freed from their stromas by treatment with collagenase retain their differentiated phenotype if they are cultured in the absence of serum. ^{9 10} However, when they are treated with 10% calf serum, these cells acquire a fibroblast phenotype over a 2- to 3-day period. ^{12 14} In serum-treated fibroblast cultures, expression of α-sm actin occurs spontaneously in a subset of cells through activation of latent TGF-βs. All keratocytes or fibroblasts in a cell culture dish can be made to acquire the α-sm actin marker by exogenous addition of active TGF-βs. ^{10 19 20} TGF-β2, produced as a paracrine factor by the corneal epithelium, recently has been shown to be the major effector of myofibroblast conversion in corneal wounds. ²¹ In contrast, the specific mediators that cause other aspects of the fibrotic transition in cornea have not been identified. 

The goal of this study was to identify physiologically relevant mediators controlling loss of the corneal crystallin TKT, using a rabbit keratocyte culture model. 

Methods

Cell Culture

Stromal cells were obtained from corneas of New Zealand White rabbits by collagenase digestion as previously described. ¹⁴ All procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Cells were plated and maintained in serum-free medium. In some experiments, cells were plated in the presence of serum, allowed to proliferate to confluence, and subcultured (serum-exposed fibroblasts), also as described. ¹⁴  

For an experiment, cells were trypsinized and replated in 35-mm dishes. Duplicate cultures were treated with supplemented calf serum (Invitrogen, Carlsbad, CA), basic fibroblast growth factor (bFGF; Sigma-Aldrich, St. Louis, MO), platelet-derived growth factor (PDGF; R&D Systems, Minneapolis, MN), or TGF-β2 (R&D Systems). The standard plating density was 2 × 10⁵ cells per dish. In experiments examining the role of cell proliferation in TKT expression, cells were plated in the range of 4 × 10⁵ to 10⁶ cells per 35-mm dish. To measure the rate of proliferation, cells were trypsinized and counted with a hemocytometer at the conclusion of the experiment. Experiments were repeated more than once to ensure reproducibility. 

Protein Electrophoresis and Immunoblot Analysis

For analysis of TKT and α-sm actin expression, cells were solubilized (30 minutes, 40°C) in PBS containing 1% Nonidet P-40 (NP-40), 0.5% sodium deoxycholate, 0.1% SDS, and protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN). Samples were then analyzed by gel electrophoresis and immunoblot. After weighing the various parameters that change with the different cell treatments planned, we decided that comparison of TKT levels on a per cell basis would give more useful information than comparison of TKT levels per unit of protein. Serum and TGF-β increase both the numbers of cells and cell size (probably increasing total protein levels per cell), whereas bFGF increases the number of cells without increasing size. Thus, the only way for us to be consistent in comparing treatments that have various effects on cell size and proliferation was to load for en equivalent number of cells. Approximately 2 × 10⁵ cells were loaded per lane for immunoblot analysis. Of particular importance, we were concerned about understanding whether loss of TKT was caused by a shut down in gene expression and dilution, due to cell replication or rapid protein degradation. We viewed these data in light of changes in cell size. We have used this method of data analysis in other studies in which cell volume changes would have confounded analysis. ^{22 23}  

For analysis, samples were subjected to SDS-PAGE. A molecular weight size standard was electrophoresed in a parallel lane. Electrophoretically separated proteins were transferred to polyvinylidene difluoride (PVDF) membrane. Amido black staining was performed to ensure that transfers were complete; if not, the blots were not used. Because TKT constitutes such a large amount of the total protein per cell, the TKT band was usually visible with the amido black stain, providing further assurance that transfer was complete. Blots were then probed with rabbit anti-serum raised against human TKT (a generous gift from Joram Piatigorsky, National Eye Institute, Bethesda, MD) at a concentration of 1:1000 or mouse monoclonal antibodies raised against α-sm actin (Sigma-Aldrich), used at 1:3000. Anti-rabbit and anti-mouse secondary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) were used at a concentration of 1:5000. Specific binding was detected by enhanced chemiluminescence (ECL; Santa Cruz Biotechnology), visualized by autoradiography, and when useful, quantified by densitometry on computer (AlphaEase imaging software; Alpha Innotech Co., San Leandro, CA). The specific band representing TKT or α-sm actin was identified by its relative electrophoretic mobility with respect to the size standard. To ensure equivalence of sample loading and transfer, protein were visualized by staining membranes with 0.1% amido black in 10% acetic acid. The standard application of duplicate treatments in the design of all experiments served to give an indication of experimental variability. 

Results

TKT Loss in Serum-Treated Keratocytes

We used immunoblot analysis to observe the change in intracellular TKT levels on a per cell basis after keratocyte exposure to serum. Representative results are shown in Figure 1 . The high-TKT level per cell observed in freshly isolated keratocytes was dramatically reduced within 2 to 3 days after culture in the presence of 10% supplemented calf serum (Fig. 1A) . (The exact time of TKT loss was somewhat variable, compare duplicates for day 2; Fig. 1A .) In contrast, the TKT level per cell remained constant when cells were plated and maintained under serum-free conditions. The extent of TKT loss was dose-dependent, as treatment of serum-free cultures with 1% serum was less effective than treatment with 10% serum (Fig. 1B) . When freshly isolated keratocytes were plated and exposed to serum for only 24 hours and then changed to serum-free conditions for four more days, TKT levels per cell were maintained (Fig. 1C) , indicating that there is a critical duration of serum exposure necessary for TKT loss. Cells cultured in 10% serum did not re-express TKT when trypsinized and replated for subculturing (serum-exposed fibroblasts). In addition, withdrawal of serum could not induce these cells to re-express TKT (Fig. 1D) . 

Compared with stromal cells plated in the presence of 10% serum, cells plated in the absence of serum undergo little proliferation. ¹⁰ Furthermore, we found that cells proliferated only marginally when plated in low concentrations of serum or given a short serum exposure (as in the experiment shown in Fig. 1 ). Therefore, we hypothesized that TKT loss might be linked with cell replication. To examine this idea, we compared TKT loss on a per cell basis in keratocytes plated at different initial densities and allowed to replicate over a 5-day period, until they attained confluence. Representative results are shown in Figure 2 . Cells plated at 50% confluence underwent one doubling, whereas cells plated at 25% confluence underwent two doublings over a 5-day period. However, TKT loss per cell occurred rapidly and with similar kinetics in both cases, with an approximate 65% decrease after the first cell division (Fig. 2A) . Comparison of rapidly proliferating cells to nonproliferating cells yielded a different result (Fig. 2B) . In cultures plated at 10% confluence, the TKT level per cell decreased by approximately 80% within 5 days, as cultures went through approximately 10 doublings. However, TKT loss per cell was only approximately 18% after 5 days in cultures plated at 100% confluence. These data suggest that serum-induced loss of TKT is associated with serum’s ability to induce cell proliferation. 

PDGF- and bFGF-Induced TKT Loss

The cytokines PDGF and bFGF are two components of serum that have been implicated in the tissue repair response, with effects on cell proliferation, morphology, and gene expression. ^{10 24} We examined the response of keratocytes to these agents over a 5-day period, in freshly isolated cells plated in the absence of serum. PDGF stimulated keratocyte proliferation with a 64% increase in number of cells over untreated controls at the highest dose (100 ng/mL). Representative results are shown in Figure 3 . PDGF-treated keratocytes also underwent a morphologic change involving an overall elongation of cell processes (Fig. 3A) , but did not spread and become spindle-shaped as did serum-treated cells. ¹⁰ The cytokine bFGF also stimulated proliferation, with a 42% increase in cell number over controls at the highest dose used (10 ng/mL). In contrast, cellular morphology was not altered (Fig. 3A) . Levels of TKT, as analyzed on a per cell basis, declined by approximately 65% in both PDGF- and bFGF-treated cultures (Figs. 3B 3C) . This finding was replicated in additional experiments. These findings again link TKT loss with cell proliferation and further show that changes in cell morphology are not essential for TKT loss. 

Effect of TGF-β2 on TKT

TGF-βs are a third component of serum implicated in the tissue repair response. ^{10 20 21} In serum-exposed fibroblast cultures, including those derived from the corneal stroma, TGF-βs induce transition to the myofibroblast phenotype (i.e., stimulate expression of α-sm actin). It has been reported that keratocytes cultured under serum-free conditions can also undergo this transition. ¹⁰ We were able to reproduce this finding, as depicted in a representative experiment shown in Figure 4 . After 5 days of TGF-β2 treatment at a dose of 1 ng/mL, keratocytes withdrew cellular processes, spread on the plate, and adopted the characteristic fibroblast spindle shape (Fig. 4A) . This resulted in an overall increase in cell size and, presumably, in total cell protein. In addition, cells began to express α-sm actin (Fig. 4B) . In contrast and despite the adoption of the fibroblast shape and α-sm actin expression, the TKT protein level, on a per cell basis, did not decrease (Fig. 4B) . In fact, the TKT levels appeared to increase, probably reflective of the increase in overall cell size with TGF-β treatment. TGF-β2 treatment also induced α-sm actin expression in serum-exposed fibroblasts that had been subcultured. However, TGF-β2 treatment did not restore TKT expression to these cells (Fig. 4B) . These results indicate that TKT loss is regulated differently from fibroblastic shape change and the α-sm actin marker. 

Cell Proliferation and TKT Loss

TGF-βs have a bimodal effect on fibroblast proliferation with the greatest effect at lower doses. TGF-βs have also been reported to stimulate cell proliferation in cultures of serum-free keratocytes. ¹⁰ To examine the relationship between cell proliferation and TKT expression in corneal keratocytes, a dose-response experiment was performed (Fig. 5) . Increasing the concentration of TGF-β2 resulted in an increase in expression of α-sm actin per cell, as assayed by immunoblot analysis. An inverse relationship between TGF-β2 dose and cell proliferation was observed: TGF-β2 at 100 pg/mL stimulated a 46% increase in number of cells over the control, whereas a dose of 10 ng/mL stimulated only a 10% increase in cell number. Despite the stronger effect on proliferation, TGF-β2 treatment at 100 pg/mL was no more effective at inducing loss of TKT than treatment at 10 ng/mL. ¹⁰ These findings indicate that the stimulation of cell proliferation is not sufficient to induce TKT loss in corneal keratocytes. 

Discussion

It has been reported that corneal crystallins are lost from resident cells of the corneal stroma during the process of wound repair, and that this loss is associated with diminished cell transparency. We and others have reported that corneal crystallins are also lost when freshly isolated corneal keratocytes are exposed to serum in culture. ^{8 25} The goal of this study was to identify factors in serum capable of inducing loss of TKT. In the current study, keratocyte cultures maintained in the absence of serum failed to lose TKT, similar to findings regarding a second corneal crystallin ALDH3A1. ⁸ Therefore, the serum-free culture model was used to screen for relevant paracrine factors. In the process of examining TKT loss in comparison to acquisition of other fibrotic markers, we show uncoupling from markers of the fibroblasts and myofibroblasts associated with fibrotic repair in the cornea and other organs. 

TKT Loss and Cellular Proliferation

In a previous publication, we showed that TKT loss is at least partially the result of a proteasome-dependent degradative process. ²⁵ Although these experiments suggest that TKT is degraded during the repair phenotype transition, declines in detectable TKT could reflect a dilution of TKT as a result of cell proliferation. If TKT transcription simply ceases on stromal cell removal from corneal tissue, the preexisting TKT protein could be passively diluted as cells divide. This would explain the retention of TKT in serum-free cultures, because they undergo little cell division. ¹⁰ However, if loss of TKT were simply a matter of dilution, protein levels would decline 50% after each successive cell division. We found that this was not the case, in that most (65%) of the decline in TKT, as measured on a per cell basis, occurred after the first cell division. This finding suggests that TKT loss is the result of both degradation and dilution. Preventing cell proliferation by plating at an initial confluent density inhibited most of the TKT loss, suggesting that TKT decline, whether by degradation or dilution, is correlated with cell cycle entrance. The association of TKT loss with cell replication was further corroborated in bFGF and PDGF cultures. However, we further showed that not all proliferation-inducing paracrine factors decrease levels of TKT. Low doses of TGF-β2 induced keratocyte proliferation, but did not cause a decline in TKT any greater than high doses, which induced little proliferation. This indicates that cell replication is not sufficient for loss of corneal crystallin. 

TKT Loss and Wound Type

In a previous study characterizing TKT loss from keratocytes during corneal repair, a freeze injury model was used. ¹⁶ This is one of a group of corneal insults applied to the ocular surface that causes the death of keratocytes, but only minimal damage to the corneal extracellular matrices. Keratocytes also die in the stroma underlying areas where the epithelium has been mechanically abraded from the basement membrane. Both injuries stimulate a regenerative type of repair response whereby stromal cells are replaced without fibrosis or scarring. In a recent study, we showed that the corneal epithelium is the primary source in corneal wounds for TGF-β, the primary stimulator of the fibrotic response. ²⁶ We further showed that TGF-β2 fails to be released into the stroma from the overlying epithelium in the abrasion injury, because the basement membrane remains intact. In this study, we show that TGF-β2 is not responsible for TKT loss. This fits with the previous finding that TKT is lost from keratocytes during regenerative repair in cornea. Our data suggest that bFGF and PDGF are likely candidates for inducing TKT loss during regenerative repair. 

Uncoupling of TKT Loss from Other Fibrotic Repair Markers

Cytokines of the TGF-β family are important mediators of the fibrotic response in vivo. In cell culture, they stimulate cell transformation to a phenotype with some characteristics of the myofibroblasts in corneal wounds characterized by extensive actin stress fibers and new expression of α-sm actin. ¹⁹ However, TGF-β2 treatment of cells cultured in serum-free medium failed to induce loss of TKT on a per cell basis. There is precedence for heterogeneous myofibroblast populations within wound regions. Skalli et al. ²⁷ showed that myofibroblasts are heterogeneous in their content of actin isoforms and intermediate proteins. Furthermore, depending on the type of scar formation (normal versus hypertrophic), the myofibroblast in the pathologic condition is slightly different and is characterized by features indicative of smooth muscle differentiation. ²⁷ It would be very interesting to learn whether the uncoupling of TKT expression from α-sm actin expression also occurs in vivo. If so, the presence or absence of TKT may create further myofibroblast heterogeneity. 

The ability of bFGF to induce proliferation while maintaining the differentiated keratocyte morphology and increasing KSPG has led to the hypothesis that bFGF is a differentiation factor for the corneal stromal cell. ^{10 28} Therefore, it is surprising that bFGF is capable of inducing loss of TKT, which is a marker of an activated cell. It is possible that during keratocyte development, the bFGF differentiation signal may occur only after keratocytes have reached a certain density within the stroma. The inability of keratocytes to proliferate would allow for maintenance of TKT levels, whereas bFGF would be capable of inducing quiescent phenotypes such as KSPG. These data suggest that keratocytes require a specific sequence of signaling events for proper differentiation to occur. 

A New Model for Repair Activation of the Corneal Keratocyte

We have proposed a model for keratocyte activation during corneal repair as a unidirectional transition in cellular phenotype, from quiescent keratocyte to activated fibroblast or myofibroblast, which cannot be reversed. ² Recent evidence suggests that this may be an oversimplification. Berryhill et al. ⁸ reported that some markers of the quiescent phenotype could be restored by withdrawing serum from stromal cell cultures; however, the corneal crystallin ALDH3A1 did not return. This is consistent with in vivo findings in the freeze wound model, in which stromal cells return to a quiescent morphology, appearing identical with keratocytes, but without expression of the crystallin TKT. 

Similar to a partially restored keratocyte phenotype, our ability to uncouple fibrotic markers from TKT loss suggests that the process of keratocyte activation in response to repair stimuli does not have to be an all-or-nothing process. The idea that wound phenotypes are differentially regulated suggests a nonlinear or multistep cellular response depending on the activating factors present during the repair process. Our findings suggest that targeting of the specific cytokines PDGF and bFGF, or the signaling pathways they activate, could enable retention of corneal crystallin in stromal cells during repair. This would make possible a more regenerative outcome for surgical procedures that result in keratocyte activation and may be a useful therapeutic strategy. 

 Supported by National Eye Institute Grants EY09828, EY13078, and EY14801; the Massachusetts Lions Eye Research Fund, Inc.; and unrestricted grants from Research to Prevent Blindness, Inc. (to Tufts University and the University of Miami). MEF was a Jules and Doris Stein Research to Prevent Blindness Professor and now holds the Walter G. Ross Chair in Ophthalmic Research.

 Submitted for publication September 24, 2003; revised May 28 and July 16, 2004; accepted August 3, 2004.

 Disclosure: B.M. Stramer, None; M.E. Fini, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

TKT was lost from keratocytes cultured with 10% serum. Keratocytes freshly isolated from the corneal stroma were plated at approximately 25% confluence and then left untreated or treated as described in each panel for 5 days. Intracellular TKT content was determined on a per cell basis by immunoblotting of samples from an equivalent number of cells at the indicated times. The band representing TKT is identified (approximately 70 kDa). The migration position of size standards is also indicated. (A) Duplicate cultures treated with 10% serum after plating were sampled daily for 5 days. Parallel cultures that were never exposed to serum were sampled at 1 and 5 days after plating for comparison. Please note that the difference between duplicate treatments is an indication of experimental variability. (B) Duplicate cultures treated with 10% serum, 1% serum, or left serum-free were sampled 5 days after plating for comparison. (C) Duplicate cultures left serum-free, treated with 10% serum (5 days serum), or treated with 10% serum for 24 hours before replacing with serum-free medium (remove serum) were sampled 5 days after plating for comparison. (D) Comparison of freshly isolated keratocytes plated and cultured in the absence of serum for 5 days (primary) with cells that had been cultured and passaged in the presence of serum and then plated and serum starved for 5 days (SEF, serum-exposed fibroblasts).

Figure 2.

 

TKT loss was associated with cell proliferation. Keratocytes were plated at different initial densities and then exposed to 10% serum or left serum free (control). Intracellular TKT content of cultures was determined after 5 days by immunoblot analysis of samples containing equivalent cell numbers. Histograms show the density of TKT bands relative to the untreated control (set at 1). (A) Keratocytes were plated at two different initial densities, such that over a 5-day period, cells divided either once (1X) or twice (2X) when exposed to serum before reaching confluence. (B) Keratocytes were plated at ∼25% confluence to allow for proliferation (low density) or at confluence so that cell division was inhibited.

Figure 3.

 

Serum cytokines PDGF and bFGF stimulated TKT loss. Freshly isolated keratocytes were plated in the absence of serum and left untreated (control) or were treated with PDGF (100 ng/mL) or bFGF (10 ng/mL) for 5 days. (A) Phase-contrast images showing changes in cell morphology after treatment. (B) Intracellular TKT content of cultures after treatment was determined by immunoblot analysis of samples containing an equivalent number of cells (top). The difference between duplicate treatments is an indication of experimental variability. Staining of the same blot with amido black (bottom) revealed a band corresponding to TKT in lanes loaded with control cell lysates, but absent in lanes loaded with treated cell lysates. In contrast, total protein in each lane is relatively unchanged with treatment. (C) Histograms shown the density of TKT bands from the blot shown in (B) relative to untreated controls (set at 1).

Figure 4.

 

TGF-β2 did not stimulate TKT loss. Freshly isolated keratocytes were plated in the absence of serum and left untreated (serum-free), or treated with TGF-β2 (1 ng/mL) for 5 days. A similar experiment (top) was performed on serum-exposed fibroblasts (SEF) that were plated and left otherwise untreated or were treated with 1 ng/mL TGF-β2 for 5 days. (A) Phase-contrast images show changes in cell morphology after treatment. (B) Relative levels of intracellular TKT (top) or α-sm-actin (bottom) in duplicate cultures (left) were determined by immunoblot analysis of samples containing equivalent numbers of cell and compared with cultures of SEFs (right). No comparison can be made between the intensity of the bands in the top and bottom panels, because they represent immunodetection with different antibodies and different times of exposure to film.

Figure 5.

 

Cell proliferation was not sufficient for TKT loss. Freshly isolated keratocytes were plated in the absence of serum and treated with TGF-β2 at 10 ng/mL, 1 ng/mL, or 100 pg/mL for 5 days. Cell numbers were counted at the end of this period. Intracellular TKT (top) and α-sm actin (middle) content of duplicate cultures was determined by immunoblot analysis of samples containing equivalent numbers of cell. The lower intensity of the α-sm actin band in the 1 ng/mL treatment compared with Figure 4B is simply reflective of the lower exposure time used in this experiment to observe the concentration dependence within the linear range of this detection method. Also, the difference between duplicate treatments is due to experimental variability, and the actual value is an average of the two. Staining of the same blot with amido black (bottom) revealed that cells treated with lower doses of TGF-β2 had less total protein per cell; however, the TKT band remained unchanged.