Abstract
The growth and bioluminescence of cells seeded in collagen and gelatin sponge matrices were compared in vitro under different conditions, and immune rejection was quantified and visualized directly in situ based on loss of bioluminescence activity. Mammalian cells expressing a Renilla luciferase complementary deoxyribonucleic acid (cDNA) were used to seed collagen and gelatin sponge matrices soaked in either polylysine or gelatin to determine optimal growth conditions in vitro. The sponges were incubated in tissue culture plates for 3 weeks and received 2, 9, or 15 injections of coelenterazine. Measurements of bioluminescence activity indicated that gelatin sponges soaked in gelatin emitted the highest levels of light emission, multiple injections of coelenterazine did not affect light emission significantly, and light emission from live cells grown in sponges could be measured qualitatively but not quantitatively. Histologic analysis of sponge matrices cultured in vitro showed that cells grew best in gelatin matrices. Visualization of subcutaneously implanted sponges in mice showed accelerated loss of light emission in immunocompetent BALB/c mice compared with immunodeficient BALB/c-scid mice, which was associated with increased cell infiltration. Our results indicate that sponge matrices carrying bioluminescent mammalian cells are a valid model system to study immune rejection in situ.
Sponge implants containing foreign living cells are rejected by the immune system similarly to an allograft, showing infiltration of leukocytes 2 and accumulation of proinflammatory molecules. 14 Immune responses to sponge implants have been evaluated traditionally using histologic examination and other methods that require removal of the matrix. An approach allowing investigators to observe rejection of implants directly in situ would provide additional information on the kinetics of rejection. Embedded cells expressing a luciferase complementary deoxyribonucleic acid (cDNA) are an attractive choice for this purpose, and have been used routinely for in situ visualization of bioluminescence activity in many different model systems. 15 Using bioluminescent cells to load sponge implants has allowed investigators to study the effects of a chosen treatment in real time. 16 We wanted to determine to what extent the bioluminescence activity of mammalian cells expressing a Renilla luciferase cDNA and grown in a sponge matrix in vitro is affected by different matrices and multiple injections of luciferase substrate and whether immune rejection of matrix-embedded cells could be observed in situ based on loss of light emission in immunocompetent mice compared with immunodeficient mice.
Materials and Methods
Plasmid Construct and Mammalian Cell Lines
A plasmid DNA carrying a humanized Renilla luciferase cDNA (phRUC) was used for stable transfection of mammalian cells and constructed by cloning a humanized cDNA coding for the Renilla reniformis luciferase from vector phRL-CMV (Promega, Madison, WI) into the Xba I site of plasmid vector pcDNA3.1 (Invitrogen, Carlsbad, CA) under transcriptional control of the cytomegalovirus promoter. Human 293A fibroblasts were transfected with plasmid phRUC using the Calcium Phosphate Transfection kit (Promega), and stable cell lines expressing humanized Renilla luciferase cDNA (hRUC) were selected using G418 selection, all according to the manufacturer's protocol. A cell line with the highest bioluminescence activity was chosen to be used in the present work.
Seeding of Sponge Matrices with Mammalian Cells
For work done in vitro, 1 cm3 sterile gelatin (GelFoam, Pharmacia & Upjohn, Kalamazoo, MI) and collagen (UltraFoam, MedChem Products, Woburn, MA) sponges were soaked overnight at room temperature in 0.4% polylysine (Sigma, St. Louis, MO), 0.1% gelatin in Dulbecco's Phosphate Buffered Saline (DPBS), or DPBS alone as the control. One hundred microliters of DPBS containing 1 × 107 293A cells stably transfected with phRUC was then injected into each sponge, and the sponges were cultured in six-well plates with 6 mL Dulbecco's Modified Eagle's Medium with 10% fetal bovine serum and antibiotics at 37°C in 5% CO2. The medium was changed every 3 days for incubation times ranging from 1 to 3 weeks.
For work done in vivo, two 1 cm3 sterile gelatin sponges soaked in 0.1% gelatin-DPBS solution overnight at room temperature were implanted subcutaneously into the ventral area of BALB/c and BALB/c-scid mice (Taconic, Hudson, NY) following an Institutional Review Board-approved surgery protocol as per National Institutes of Health guidelines. Twenty-four hours after surgery, one of the two implanted sponges received an injection of 100 μL DPBS containing 1 × 106 293A cells that expressed hRUC in a stable manner. The other sponge received 293A cells not expressing hRUC as the control.
Assay of Luciferase Activity
For work done in vitro, Renilla luciferase activity was measured from intact sponge matrices after injection of 100 μL of 1 μM coelenterazine (NanoLight, Pinetop, AZ) in DPBS and 0.5% ethanol into each sponge. The sponges were immediately placed into a glass scintillation vial, and light emission was measured in a Turner TD-20e or 20/20n luminometer (Promega) for 10 seconds. Culture media of sponges were also assayed for bioluminescence activity to evaluate the extent of cell release by the different sponge matrices. Media were centrifuged at 200g for 15 minutes at 4°C, supernatants were collected, and 500 μL was used for assay of luciferase activity. In addition, light emission was measured from lysates of cells grown in sponge matrices to evaluate total bioluminescence activity present in sponges. Each sponge was placed in liquid nitrogen and ground in 500 μL luciferase lysis buffer (Promega). One milliliter of Renilla luciferase assay buffer (0.5 M NaCl, 1 mM ethylenediaminetetraacetic acid, 100 mM KH2PO4, pH 7.4) was added to the lysates, which were then centrifuged at 15,000 rpm for 1 minute. One hundred microliters of supernatant was assayed for bioluminescence activity in a Turner 20/20n luminometer after injection of 500 μL of 1 μM coelenterazine.
For work done in vivo, bioluminescence activity from sponges implanted into immunocompetent BALB/C and immunodeficient BALB/c-scid mice was measured daily over 7 days or only on day 0 and day 7 postimplantation after injection of 100 μL of 10 μM coelenterazine substrate in phosphate buffered saline (PBS)-ethanol. Bioluminescent measurements were performed with an Argus 50 Hamamatsu Low Light Imaging System (Bridgewater, NJ) using Metamorph software (Molecular Devices Corporation, Sunnyvale, CA). Imaging was performed over a 10-minute integration period.
Histologic Analysis
Sponges were collected and fixed with 10% buffered formalin. Sponge sections were stained with hematoxylin and counterstained with eosin. Histologic slides were examined under a bright-field microscope using 400× magnification, and photographs were taken with a SPOT Imaging System (Diagnostic Instruments, Sterling Heights, MI).
Statistical Analysis
The results were analyzed using the Mann-Whitney U test, and p values ⩽ .05 were considered significant.
Results
In Vitro Experiments
Effects of Sponge Matrix Composition on Bioluminescence Activity of Seeded Cells
Two biodegradable sponge materials that often serve as matrices for cell implantation were used in this study, collagen and gelatin. The matrices were soaked in polylysine or gelatin solutions, which are known to promote cell attachment and growth. 17 The matrices were seeded with bioluminescent fibroblasts and incubated in culture medium for 1 week. Light emission activity was then measured directly from sponges and culture media. The results indicated that cells showed the highest levels of light emission when grown in gelatin sponges (Figure 1A). Gel-Gel sponges showed significantly higher light emission activity compared with gelatin sponges soaked in PBS and collagen sponges soaked in polylysine, whereas gelatin sponges soaked in polylysine showed significantly higher light emission activity only when compared with the collagen sponge matrix. In contrast, culture media from collagen sponges showed significantly higher light emission activity compared with all gelatin sponges (Figure 1B).
Effects of Incubation Time and Frequency of Coelenterazine Injection on Bioluminescence Activity of Cells Grown in Sponge Matrices
Sponge matrices seeded with fibroblasts synthesizing Renilla luciferase and incubated in culture medium for 3 weeks received either 2, 9, or 15 injections of a coelenterazine solution following the schedule shown in Figure 2. Sponges were assayed for bioluminescence activity at the end of weeks 1, 2, and 3. The results indicated that cells grown in collagen sponges did not show a significant decrease in light emission after 3 weeks of incubation regardless of the number of substrate injections (Figure 3). In contrast, most of the cells grown in gelatin sponges showed a significant decrease in bioluminescence activity at week 3 compared with week 1. This decrease was not dependent on the number of substrate injections because it was observed with gelatin sponges receiving 2, 9, or 15 injections of the substrate.
Effects of Sponge Matrix Composition on Cell Growth
The results from histologic analysis of sponges that received injection of fibroblasts and were grown for 3 weeks in tissue culture medium indicated that cells grew best in Gel-Gel sponges (Figure 4). In all cases, cell growth was restricted to a section adjacent to the surface of the sponges.
Bioluminescence activity of cells seeded in different sponge matrices cultured in vitro for 1 week. One hundred microliters of 1 × 108/mL human 293A cells stably expressing a humanized Renilla luciferase cDNA (hRUC) were injected into sponge matrices, which were then incubated in cell culture medium. After 1 week incubation, sponges (A) and cell culture media (B) were assayed for luciferase activity. Col-Pol = collagen sponges soaked in polylysine; Gel-Gel = gelatin sponge soaked in gelatin; Gel-PBS = gelatin sponges soaked in Dulbecco's Phosphate Buffered Saline; Gel-Pol = gelatin sponges soaked in polylysine; RLU = relative light units.
Comparison of Bioluminescence Activity from Live Cells and Lysates of Cells Grown in Gelatin Sponges over 3 Weeks
Histologic analysis of sponges incubated for 3 weeks showed that cells seeded in Gel-Gel sponges grew at a higher density compared with other sponges (see Figure 4). However, the growth increase was not associated with a comparable increase in bioluminescence activity measured at week 3 (see Figure 3). To determine whether this result was due to a limiting factor in the assay of light emitted by live cells grown in sponges, cells were seeded in Gel-Pol and Gel-Gel sponges and incubated for 3 weeks, and bioluminescence activity was measured weekly from intact (live cells) and ground (cell lysates) sponges. The results indicated that bioluminescence activity from intact Gel-Pol and Gel-Gel sponges was similar, was highest at time 0, and decreased significantly during the first week (Figure 5A). Bioluminescence activity from Gel-Gel and Gel-Pol ground sponges was also similar but, in contrast, increased progressively after week 1 and was highest at the end of week 3 (Figure 5B). Therefore, measurement of bioluminescence activity from live cells grown in sponges did not reflect an increase in luciferase activity that occured after the first week of growth.
Schedule of coelenterazine delivery and measurement of bioluminescence activity in vitro. One hundred microliters of 1 × 108/mL human 293A cells expressing humanized Renilla luciferase cDNA was injected into sponge matrices, which were then incubated in cell culture medium for 3 weeks. Sponges received 2, 9, or 15 injections (one injection on a given day) of a solution of 1 mM coelenterazine (black and white arrows), and bioluminescence activity was measured at the end of each week (white arrows). The results are shown in Figure 3.
In Vivo Experiments
Bioluminescence Activity from Cells Seeded in Sponge Matrices Implanted into BALB/c and BALB/c-scid Mice
Based on results obtained in vitro, the Gel-Gel sponge was chosen as a matrix to grow cells after subcutaneous implantation into mice. BALB/c-scid mice, which are deficient in T and B cells, and immunocompetent BALB/c mice received the sponge implants, and fibroblasts synthesizing Renilla luciferase were then injected into the sponges. Light emission from mice was quantified (Figure 6A) and visualized (Figure 6B) directly in situ over a period of 7 days, which is sufficient to observe acute rejection in animals. The results indicated that light emission decreased rapidly in both BALB/c and BALB/c-scid mice after 24 hours. After 6 days, bioluminescence activity was reduced to background levels in BALB/c mice. In contrast, BALB/c-scid mice still showed above-background light emission that could be detected 7 days after implantation. In addition, light emission activity after daily injection of coelenterazine was similar to that observed after injection of substrate on days 0 and 7. Histologic examination of gelatin sponge matrices 7 days after implantation showed increased cell infiltration into sponges implanted into BALB/c mice compared with sponges implanted into BALB/c-scid mice (Figure 7).
Effects of frequency of coelenterazine injection and incubation time on bioluminescence activity of cells seeded in different sponge matrices. Sponges were treated and assayed as described in Figure 2. Data represent the average of two measurements made from three sponges (± SD). Col-Pol = collagen sponges soaked in polylysine; Gel-Gel = gelatin sponge soaked in gelatin; Gel-PBS = gelatin sponges soaked in Dulbecco's Phosphate Buffered Saline; Gel-Pol = gelatin sponges soaked in polylysine; RLU = relative light units.
Histologic analysis of sponge matrices incubated in vitro. Sponges seeded with 293A cells synthesizing luciferase and incubated for 3 weeks were stained with hematoxylin and eosin (X400 original magnification). Col-Gel = collagen sponge coated in gelatin solution; Col-Pol = collagen sponge coated in polylysine solution; Gel-Gel = gelatin sponge coated in gelatin solution; Gel-Pol = gelatin sponge coated in polylysine solution.
Discussion
In the work presented here, we used mammalian fibroblasts expressing a cDNA coding for the Renilla reniformis luciferase to evaluate the bioluminescence activity of cells grown in sponge matrices in vitro and in vivo. We used Renilla luciferase because of our extensive experience with this reporter protein.18–20 Our goals were to determine the effects of matrix composition on bioluminescence activity in vitro-whether multiple injections of coelenterazine substrate affect the bioluminescence activity of sponge-grown cells, whether bioluminescence activity from live cells grown in sponge matrices is quantitative, and whether loss of bioluminescence activity can be used as an indicator of immune rejection in vivo.
Measurements of light emission after culturing collagen and gelatin sponge matrices seeded with bioluminescent mammalian cells showed that cells grown in gelatin sponges emitted the highest levels of light after 1 week. The sponges were also soaked in solutions containing molecules known to favor cell attachment and/or growth, that is, polylysine and gelatin as reported here, as well as fibronectin and bovine serum albumin (data not shown). Among the different sponges, Gel-Gel sponges generally showed the highest light emission activity, and cell culture media from collagen sponges showed the highest light emission activity. Altogether, the data indicated that cells attached more efficiently to the gelatin matrix compared with collagen, suggesting that collagen sponge matrices are better suited for applications in which the release of cells into the implant environment is beneficial, for example, for tissue remodeling. Based on our results, gelatin sponges may be more appropriate for the engineering of artificial tissues and organs for which cell retention is more desirable.
Bioluminescence activity of live cells and lysates of cells grown in gelatin sponge matrices for 3 weeks in vitro. Gelatin sponges soaked in gelatin (Gel-Gel) or polylysine (Gel-Pol) were seeded with 293A cells synthesizing luciferase and incubated for 3 weeks. Bioluminescence activity was measured immediately after injection of cells (0) and at the end of each week (1-3). Data represent the average of two measurements made from three sponges (± SD). RLU = relative light units.
Bioluminescence activity of sponge matrices in immunodeficient and immunocompetent mice in situ. A, Gelatin sponges coated in gelatin solution (Gel-Gel) were implanted into immunodeficient (BALB/c-scid) and immunocompetent (BALB/c) mice (2 sponges/mouse). For each mouse, one of the sponges received an injection of 1 × 106 human 293A cells expressing humanized Renilla luciferase cDNA and the other received the same number of untransfected 293A cells. Bioluminescence activity was measured immediately after cell injection (day and week 0) and every 24 hours (days 1-7) for some sponges or after 1 week for others. RLU = relative light units, ± SD. B, Implanted Gel-Gel sponges were visualized after injection of coelenterazine immediately after injection of cells synthesizing Renilla luciferase (day 0) and over the following 7 days.
Next, we wanted to determine whether multiple delivery of coelenterazine into sponge matrices would affect the bioluminescence activity of seeded cells, possibly causing changes in the light-emitting properties of mammalian cells as a result of the toxic and antioxidant properties of the substrate. 21 In addition, we also wanted to investigate changes in bioluminescence activity in vitro over a longer time period, that is, 3 weeks. The results indicated that 15 daily injections of coelenterazine did not cause a significant decrease or increase in bioluminescence activity after a 3-week incubation time compared with 2 and 9 injections. Therefore, the number of injections of the substrate did not appear to have an effect on the bioluminescence of cells over time. Nevertheless, the results indicated that, in most cases, cells grown in gelatin matrix showed a significant decrease in light emission after incubation for 3 weeks compared with the 1-week incubation, which was not observed with cells grown in collagen sponges.
Histologic analysis of implanted sponge matrices. Gelatin sponges coated in gelatin solution that received 293A cells expressing humanized Renilla luciferase cDNA (293A/RUC) or untransfected cells (293A) were taken at day 7 after subcutaneous implantation and stained with hematoxylin and eosin (x400 original magnification).
Histologic examination of the sponge matrices at the end of week 3 showed that cells had grown to higher densities in gelatin sponges compared with collagen sponges. Moreover, cells showed increased growth in Gel-Gel sponges compared with gelatin sponges soaked in polylysine (Gel-Pol sponges). However, at week 3, light emission from cells grown in Gel-Gel sponges compared with other sponges and in Gel-Pol sponges compared with similarly treated collagen sponges did not reflect the increased growth. The absence of increased bioluminescence activity with increased cell density in sponge matrices could have been the result of a direct loss of luciferase activity, the inability to detect all luciferase activity, or both. Measuring bioluminescence activity from ground Gel-Gel and Gel-Pol sponges that had been cultured in vitro for 3 weeks, that is, from cell lysates, revealed a direct loss of luciferase activity with increased cell density because there were no statistical differences in bioluminescence activity between Gel-Gel and Gel-Pol sponges at all measurement time points. The loss of luciferase activity with increased cell density could have been the result of decreased transcription, translation, and cell death.
The results also indicated that not all bioluminescence activity could be detected from intact sponges because the activity remained relatively constant at the end of each week when measured from intact sponges, but it increased steadily when measured from ground sponges. The inability to detect all bioluminescence activity from intact sponges could have been the result of lower availability of oxygen and coelenterazine, which are both required for catalysis of light emission by Renilla luciferase. However, the amounts of coelenterazine injected into intact sponges were not limiting because bioluminescence activities at the end of weeks 1 to 3 were significantly lower than at time 0, indicating that the limiting factor was dependent on the growth condition. Altogether, the results indicated that measurement of bioluminescence activity from live cells grown in sponge matrices was not quantitative but qualitative.
Because Gel-Gel sponges showed the highest bioluminescence activity after 1 week incubation in vitro, they were chosen for subcutaneous implantation into BALB/c and BALB/c-scid mice to determine the possible effects of the immune system on bioluminescence of seeded cells over a 1-week period. The results showed that bioluminescence activity decreased 20-fold over 24 hours after injection of 293A cells expressing hRUC into implanted sponges. The decrease was likely due to an intrinsic loss of cell bioluminescence because it was observed in both immunocompetent and immunodeficient mice. Nevertheless, over the course of the next 6 days, bioluminescent cell-carrying sponges implanted into immunocompetent BALB/c mice showed an accelerated loss of light emission compared with similar sponges implanted into BALB/c-scid mice. By day 6, only background bioluminescence could be detected from sponges in BALB/c mice, which showed heavy cell infiltration, as reported previously by others.2,22,23 In contrast, sponges implanted into BALB/c-scid mice were still emitting light and showed little cell infiltration. Altogether, these results indicated that the loss of light emission in BALB/c mice was due to an immune response.
In conclusion, our data indicate that sponge matrices seeded with bioluminescent mammalian cells are a valid model system to follow rejection of cells in situ. Sponge matrices offer the possibility to select the type of cell to be seeded and to engineer or treat cells in different ways. Because of its flexibility and simplicity, this model system is a useful addition to the use of transgenic mice expressing luciferase to study immune rejection. 24