- Email: [email protected]
The effect of three Korean traditional medicines on the growth rate of cultured human keratinocytes
Journal of Ethnopharmacology 74 (2001) 53 – 61 www.elsevier.com/locate/jethpharm
The effect of three Korean traditional medicines on the growth rate of cultured human keratinocytes Seok Hee Chung a,1, Hiroto Terashi a, Lenore M. Rhodes a, Namdoo Moon b, William R. Dunham b,*, Cynthia L. Marcelo a a
Di6ision of Plastic and Reconstructi6e Surgery, Department of Surgery, Uni6ersity of Michigan Medical Centers, Ann Arbor, MI 48109, USA b Biophysics Research Di6ision, The Uni6ersity of Michigan, 930 North Uni6ersity A6enue, Ann Arbor, MI 48109 -1055, USA Received 1 July 2000; received in revised form 14 August 2000; accepted 21 August 2000
Abstract The effect of three different Korean Traditional Medicines (KTM) was studied on several functional parameters of adult human cells in culture. The cells were non-transformed strains of normal, skin epidermal cells (keratinocytes) from adult humans. Aqueous extracts of the herbal medicines were tested using two types of cell strains: one type was essential fatty acid deficient (EFAD) cells which grow rapidly in medium that was low in calcium and had no essential fatty acids; the second type was a cell strain grown in medium supplemented with essential fatty acid (EFA-supplemented). These cells had much slower, in vivo skin growth rates, and the fatty acid composition resembled that measured in epidermal biopsy tissue. The KTMs chosen for this study were tae-gang-hual-tang (for treating osteoarthritis), hual-ak-tang (for pain relief) and sip-zeon-tae-bo-tang (for fortifying immune systems). Because high proliferation rates usually correlate with skin inflammation and because many of the chemotactic agents mediating inflammatory response are modified fatty acids, this study focused on cell growth rate and membrane fatty acid composition as signals for the effects of the herbal medicines. By monitoring growth rate, these experiments measured both a stimulatory and a regulatory effect on the growth of keratinocytes. Some toxicity was seen at the highest doses of the KTMs. These effects were modeled mathematically, and the results showed varying effects on growth rate depending on dose and herbal recipe. The fitting parameters were discussed as they relate to biological function. The experimental design was also discussed and alternatives were suggested. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Epidermal cells; Epidermis; Gas chromatography; Alternative medicine; Mathematical model; Adult human cell strains
* Corresponding author. Tel.: + 1-734-7636721; fax: + 1-734-7643323. E-mail address: [email protected] (W.R. Dunham). 1 Present address: Department of Physiotherapy, College of Oriental Medicine, Kyung Hee University, Seoul 130-702, South Korea. 0378-8741/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 8 7 4 1 ( 0 0 ) 0 0 3 4 2 - 1
54
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
1. Introduction The medicines used in Korean Traditional Medicine (KTM) consist of appropriate amounts of mixed natural products, such as herbs, animal derivatives, and minerals. Each component is measured according to the principles of KTM theory, which considers synergistic effects and drug interactions. Typical KTMs have five to fifteen components, which interact to mollify any toxicity, to minimize side effects and to maximize the effect of the drug. In most cases, KTMs are mixed with water, boiled and extracted yielding a tea, which is drunk by the patient (Hur, 1989; Doo, 1997). One possible approach to studying these medicines is to identify their active ingredients and to test each one until a correlation is discovered between an ingredient and the curative effect of the medicine. We rejected this method of study for several reasons. Experience has shown that the three KTMs are valid healing agents and that all of the herbs and active ingredients are necessary for the medicine to be effective. Therefore, it is not possible to study only the effect of a component of a specific KTM; the entire recipe must be kept intact. Secondly, drug companies long ago have mined oriental medicines for their ingredients. Lists of the known active ingredients of these herbs are available in the references to this manuscript and elsewhere. These lists are long; to study each would not be possible in the time frame for this study. Furthermore, it is probable that this type of study would reproduce work already done in drug companies and, in view of the need to retain all of the KTM ingredients; we felt that this type of study would have been unlikely to produce new and useful information. Oriental medicines are effective treatments in most diseases and are most effective in the treatment of functional (no visible anatomical effects) or chronic diseases. Standard practice of KTM prescribes that mechanical or acute diseases are treated more effectively by the modern technology of Western medicine. In Korea, KTM treatment is most prevalent in the management of chronic illnesses in the elderly, which comprise a large portion of the ailing population. Even though
widely practiced, a fundamental, scientific understanding of the biochemical mechanism of any often-used KTM formulation is not available to the medical community. To simplify and increase the efficacy of KTM, it would be beneficial to develop standard methods to identify the main classes of active drugs in KTM. Thus far, all KTM drug studies have been performed using animal models. We therefore decided to develop methods to test the active drugs of KTM using a primary human epidermal cell culture system. It was decided to use this cell system because it is a cell strain whose life span has not been transformed by modification of its DNA. Accordingly, the cell strain has a limited life span, and these cells are grown under very well-defined medium conditions. The adult human epidermal cells are routinely used to create both temporary and permanent skin grafts for wound healing, forming ‘normal-appearing’ tissue (Pittelkow and Scott, 1986; Langdon et al., 1988; Falanga et al., 1998). The epidermal cells are harvested using the enzyme, trypsin, from split thickness skin that is routinely discarded from elective surgical procedure at the University of Michigan Medical Center. The resulting primary basal epidermal cell cultures are keratinocyte cultures. Growth of keratinocytes in serum- and lipid-free, low calcium conditions results in rapidly growing cultures that can be greatly expanded by numerous subculturing procedures (passages), without the loss of viability (Boyce and Ham, 1985a,b). The cultures grow well to the sixth passage when senescent (non-dividing) cells become the dominant cell type in the culture (Marcelo et al., 1992). Keratinocytes grown under these conditions are severely deficient in essential fatty acids (EFA) by the first passage (Marcelo et al., 1992). These cells have greatly diminished amounts of the major n-6 polyunsaturated fatty acids (18:2-, 20:3- and 20:4fatty acids), and contain increased amounts of the monounsaturated fatty acids, 16:1(n-7) and 18:1(n-9). Using published methods to control the fatty acid composition of normal adult human keratinocytes in culture (Marcelo et al., 1994), the fatty acid composition of the keratinocyte cultures were normalized to the levels found in human epider-
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
mal cell membranes. The fatty acid restored cultures presented a slowed growth rate, a decrease in successful passage and a more visually differentiated (flat and cornified) cellular appearance (Marcelo et al., 1992). Analysis of keratin type and cell envelope formation using PAGE and monoclonal antibodies led to the conclusion that the cells senesced more rapidly, rather than differentiated, under the influence of the EFA family of lipids (Cohen et al., 1995). This manuscript describes a study of the effects of three different KTMs on the growth rate and other physiologic parameters of human keratinocytes grown in both the EFA-deficient and -supplemented media. The KTMs were selected from 320 herb extract formulas commonly used at the Kyung Hee University Medical Center, Seoul, Korea: Tae-gang-hual-tang (HH 16) (Hur, 1989; Doo, 1997) is used for osteoarthritis, hual-ak-tang (HH 140) Hur, 1989 for low back pain, shoulder pain, and sciatica, and sip-zeon-tae-bo-tang (HH 76) (Hur, 1989; Doo, 1997) for fortifying the immune system. All of these medicines are expected to change the inflammatory response of the affected tissue. Because there is a large body of evidence concerning the effects of various reagents on skin growth, it was decided to use human epidermal keratinocytes as a platform for this study. The focus of our laboratory is the involvement of fatty acids in this response. Combining the interest in fatty acids with our experience with this keratinocyte culture technique led to the design of the experiments. For example, it has been found that cell proliferation and cell growth rate are often correlated in the human skin disease, psoriasis (Wright, 1983). Accordingly, the effect of these medicines on keratinocyte number, total protein and DNA, and cell fatty acid composition was investigated as gauge of the effect of these three KTMs on human tissue in general.
2. Material and methods
2.1. Keratinocyte culture The basic medium, MCDB 153, was prepared as previously described (Boyce and Ham, 1983,
55
1985a,b) and was supplemented with 0.6 mM (0.218 mg/ml) hydrocortisone, 5 ng/ml epidermal growth factor (EGF), 5 mg/ml insulin, 6 mg% bovine pituitary extract (3ml of extract, 10 mg protein/ml, in 500ml of medium) and 0.15 mM CaCl2 to form complete growth medium. All components of the medium, including the bovine pituitary extract, are lipid-free. Discarded human skin from mastectomy and abdominoplasty cosmetic surgery was rinsed and soaked for 2 h in solution A (30 mM hydroxyethypiprazine-N%-2-ethansulfonic acid (HEPES), 10 mM glucose, 3 mM KCl, 130mM NaCl, 1.0 mM Na2HPO4, pH 7.4) (Boyce and Ham, 1983). The skin was cut into 0.2–0.3 × 7 cm strips and digested overnight at room temperature with 0.08% trypsin in solution A, which separates the skin at the dermal-epidermal junction, leaving the basal cells on the dermal layer. The basal epidermal cells (keratinocyte) were gently scraped from the separated dermis into MCDB 153 plus 10% PBS. The filtered cell suspension (250 mm mesh, PGC Scientific, Gaithersburg, MD) was plated at 20 million cells in 15 ml MCDB 153 containing 2% chelexed FBS per T-75 flask and was grown at 37°C in 5% CO2 in air gassing. The cultures were fed every 48 h with serum-free MCDB 153 and reached approx. 70% confluence after 3–5 days (Boyce and Ham, 1983, 1985a,b; Willie et al., 1984). Usually, the monolayers were passaged every 3rd to 5th day, yielding an expanding strain from each surgical specimen. 2.2. Fatty acid supplementation The stock fatty acids (18:2 (n-6), 16:0 and 20:4 (n-6)) were stored in chloroform (Marcelo et al., 1994). A 30× final concentration (10 mM 18:2 and 5 mM 16:0 and 20:4 fatty acids) was evaporated under an N2 stream (sterilized using a Millex–SLGS, Millipore filter Millipore, Boston, MA). Ten microliters of 95% ethanol was added to the tube to solubilize the fatty acid, plus 1/30th final volume of sterile 0.6% bovine serum album in medium. After vigorous vortexing, the fatty acids plus carrier were added to the medium. The final amount of bovine serum albumin was 0.02%. The medium was freshly prepared at each change of medium.
56
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
2.3. Herb preparation and extract supplementation KTMs were prepared by the Kyung Hee University oriental medical hospital. Each prescription is shown in the following section. All herbs were dried and weighed. The prescription was added to a 10× weight of water, boiled for 1.5 h, and extracted. The extracted tea was then filtered using a centrifuge, concentrated by vacuum falling filter evaporator, dried by spray dryer, and granulated (Doo, 1997). Specifically, 6 g of herb extract was dissolved in medium to a final concentration of 0.6 g/10 ml for 1 h at 37°C. It was then filtered-sterilized and added to the MCDB 153 medium. The keratinocytes were fed with this medium every 48 h. Cells were grown in both the control (EFA-deficient) and fatty acid supplemented (EFA-supplemented) media.
2.4. Herb formulae 2.4.1. HH16(Tae-gang-hual-tang) Ingredients by descending order of weight: Notopterygium incisum, Cimicifuga heracleifolia, Angelica pubescens f. biserrata (6 g each); Akebia quinata, Chaenomeles sinensis, Achyranthes bidentata, Areca catechu (4 g each); Atractylodes japonica, Atractlodes macrocephala, Stephania tetrandra, Clematis manshurica, Angelica gigas, Poria cocos, Crataegus pinnatifida var. major, Hordeum 6ulgare var. hexastichon, Alisma plantago-aquatica var. orientale, Glycyrrhiza uralensis (3 g each). 2.4.2. HH140 (Hual-ak-tang) Ingredients by descending order of weight: Chaenomeles sinensis, Chelidonium majus (10 g each), Corydalis turtschaninovi (8 g), Clematis manshurica, Angelica pubescens f. biserrata, Angelica gigas, Rehmannia glutinosa, Paeonia lactiflora, Atractylodes japonica (6 g each); Citrus unshiu, Boswellia carterii, Commiphora myrrha, Zingiber officinale, Zizyphus jujuba var. inermis (4 g each); Carthamus tinctorius, Amomum villosum (3 g each); Glycyrrhiza uralensis (2 g).
2.4.3. HH 76 (Sip-zeon-tae-bo-tang) Ingredients by descending order of weight: Panax ginseng, Atractylodes macrocephala, Poria cocos, Rehmannia glutinosa, Paeonia lactiflora, Cnidium officinale, Angelica gigas (6 g each); Glycyrrhiza uralensis, Astragalus membranaceus, Cinnamomum cassia (4 g each). 2.5. Analysis of fatty acid methyl esters For lipids analysis, the cell monolayers were drained of medium, rinsed twice with cold phosphate buffered saline (PBS). Two milliliters of HPLC grade methanol was added over the cell layer. After scraping the monolayer, the material was then extracted with a 1:2:1.5 ratio of methanol:chloroform:0.1M KCl in 50% methanol, and the organic phase was processed as previously reported (Marcelo et al., 1992). The sample was evaporated under nitrogen stream, re-suspended in 50 l of 1:1 chloroform:methanol, applied to a thin-layer chromatography (TLC) silica gel 60 plate (Merck, Darmstadt, Germany) and chromatographed in one direction using CHCl3:MeOH:acetic acid (90:8:1). All the phospholipids, which remained at the origin, were scraped and the lipids eluted from the silica during transmethylation with 6% methanolic–HCl+ 50 mg 17:0 as an internal, quantitative standard. The fatty acid methyl esters were extracted into petroleum ether (Baker, Phillipsburg, NJ), evaporated to dryness and stored frozen under benzene. For analysis, the samples were evaporated to 200 ml of benzene, filtered using a 0.45 mm filter, were evaporated and re-suspended in 100–150 ml filtered chloroform. One microliter chloroform was injected into the gas chromatograph for analysis. The samples were analyzed using a Hewlett Packard (Wilmington, DE) GC Model 5710A equipped with a J & W Scientific (Folsom, CA) fused silica Megabore DB225, 0.53-m diameter column. A 21-fatty acid standard (Nu-CHEKPREP, Inc., Elysian, MN) was used to calibrate the GC elution times and to identify the fatty acids. A known amount of 17:0-fatty acid was added as an internal standard, because its peak is easily identifiable and quantified when the chro-
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
57
matogram is digitized by an IBM-PC computer interface (Model AN-146; Alpha Products, Darien, CT). We wrote the recording and data evaluation software in BASIC or FORTRAN. The data are presented as mg% or as mg fatty acid per mg of protein.
T-distribution (standard techniques). The parameter values in the tables are the means (N= 3).
2.6. Cell proliferation studies
3.1. Fatty acid composition
For studying cell proliferation, the cultures were lightly trypsinized and counted using a hemocytometer; trypan blue exclusion was used to determine cell viability. DNA Burton, (1968) and protein (Lowry et al., 1951) assays were used to quantitate these materials. The control cells (no KTM) were allowed to grow to 70-80% confluence (3-5 days); then both the control- and KTMtreated cell numbers were determined.
Fatty acid composition of the cellular lipids was monitored by gas chromatography (GC). For each passage, the lipids of the control and KTMtreated cells were measured and compared. In no case was there a change in the fatty acid composition due to the addition of the KTMs to the cell growth medium. Our typical results are shown in Fig. 1, where the GC chromatogram of P3 (passage three) EFA-deficient control cells is plotted. Compared to these control cells, the EFA-supplemented cells presented a greater 18:2-fatty acid peak (7.8 min), a smaller 16:1-fatty acid peak (4.5 min), a smaller 18:1-fatty acid (7.1 min) peak and much greater 20:3- and 20:4-peaks (13.9 and 14.4 min).
2.7. Statistical analysis Each experiment was performed three times, each with a different cell strain (from a different primary culture). As we will show below, the data were analyzed using curve-fitting procedures to yield three parameters for each experiment: alpha, beta and gamma. These values (N =3) were analyzed to provide information on statistical significance of the values in the tables using the
Fig. 1. GC spectra of the fatty acids in EFA-deficient keratinocytes. The elution times are: 2.9 min= 14:0; 4.2 min= 16:0; 4.5 min= 16:1; 5.3 min =17:0 (50 mg internal standard); 6.7 min= 18:0; 7.1 min =18:1; 7.8 min = 18:2; 12.0 min = 20:1; 14.4 min =20:4; 18.3 min = 22:0; 27.5 min = 24:0; and 28.9 min= 24:1 fatty acid. The identifications are made using the calibration standard shown in the text.
3. Results
3.2. Cell growth rates Cell growth as a function of drug dose was determined by measuring the total number of cells in the cultures, the amount of DNA in the cultures, and the amount of protein in the cultures. All measurements were converted to ‘% control’ values by dividing the measurements by the appropriate control (no drug) and multiplying by 100. The data from one of the dose/response tests of the KTM (HH16) cells are shown in Fig. 2. These data are representative of all of the results for all the combinations of drugs and cells. At the lowest drug additions to the media (both EFAdeficient and supplemented), the growth numbers were larger than 100, indicating an enhancement of the growth numbers compared to the control values. At higher dose values, this enhancement effect decreased with increasing dose values. In some cases, the cell population as well as the DNA and protein values decreased to a value far smaller than 100, which indicated that the drug was toxic to the cell culture.
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
58
Fig. 2. Modeling of the Effects of KTM on keratinocytes in culture. Cell growth as a function of drug dose was determined by measuring the total number of cells in the cultures, expressed as ‘% control’. In this figure is presented the data from one of the dose/response tests of the KTM (HH16) cells. These data are representative of all of the results for all the combinations of drugs and cells. Table 1 Cell number, DNA, and protein fit parameters [Eq. (Eq. (1))] for the cells grown in EFA-deficient and supplemented medium and HH16 KTM. P I, P2 and P3 refer to consecutive passages of the cell strains, P I being the first passage after the primary culture of the keratinocytes a
P1 P2 P3 P1 P2 P3 P1 P2 P3 P1 P2 P3
b
EFA-deficient Cell number 0.041 0.015 0.047 0.018 0.042 0.029 EFA-deficient DNA 0.012 0.025 0.024 0.031 0.078 0.072 EFA-deficient Protein 0.010 0.022 0.030 0.011 0.038 0.030 EFA-supplemented Cell Number 0.014 0.015 0.026 0.026 0.062 0.040
g
0.004 0.004 0.004 0.001 0.001 0.001 0.001 0.011 0.001 0.005 0.018 0.038
We have mathematically modeled these two effects by the following equation, G = 100axe-bx + 100e-gx
(1)
where G signifies the cell growth (cell c , DNA or protein level); x signifies the dose added to the medium (ml); and a, b and g are data fitting parameters. Both terms in the above expression contain exponential functions, which result mathematically from the assumption that the attenuation of the relevant term is proportional to the dose. In the first term, the enhancement of cell growth, ax, is attenuated by a control argument, bx. If a is equal to zero, then the size of b is irrelevant because there is no enhancement to control. If the toxicity parameter, g, is equal to zero, then the value of G remains equal to 100 when the first term becomes negligible at very large doses. A larger value for g indicates a higher toxicity. In Table 1, Table 2, and Table 3 are listed measured values of the fitting parameters at passage 1, 2, and 3 for each of the drug compositions, HH16, HH140 and HH76. In all cases except for the HH16 drug, the cells were grown in medium with no fatty acid. For the HH16 drug tests, the cells were also grown in fatty acid enhanced medium as indicated in Table 1.
3.3. The effect of KTM HH16 on cultured keratinocytes (Table 1) For the cell cultures grown without fatty acids in the culture medium, EFA-deficient, g was not significantly different from zero; in other words, Table 2 Cell number, DNA, and protein fit parameters [Eq. (1)- for the cells grown inEFA-deficient medium and HH140 KTM a
P1 P2 P3
0.009 0.027 0.098
P1 P2 P3
0.009 0.027 0.079
P1 P2 P3
0.008 0.030 0.072
b Cell number 0.008 0.027 0.030 DNA 0.008 0.027 0.035 Protein 0.008 0.013 0.025
g
0.007 0.008 0.030 0.008 0.008 0.009 0.008 0.013 0.025
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61 Table 3 Cell number, DNA, and protein fit parameters [Eq. (Eq. (1))] for the HH 176 EFA deficient cells a
P1 P2 P3
0.014 0.059 0.041
P1 P2 P3
0.000 0.118 0.127
P1 P2 P3
0.041 0.037 0.058
b Cell number 0.000 0.026 0.033 DNA 0.108 0.058 0.048 Protein 0.033 0.016 0.024
59
toxic to the cell cultures at the highest passage number compared to the results from the test of the other two drug recipes, HH16 and HH140.
g
0.019 0.000 0.033 0.000 0.000 0.048 0.008 0.000 0.023
there was no significant toxicity from KTM HH16 for these cells. For the EFA-deficient (highly proliferative) cells, the value of a (cell number) was constant at an average value of 0.044 90.2 and the value of b was fairly constant at 0.0219 0.005 (average9 SEM); however, the DNA and protein values showed a tendency to increase with passage number. This tendency was also seen in the EFA-supplemented (normal growth rate and fatty acid composition) cells, where the value of g also showed this tendency to increase with passage number.
3.4. The effect of KTM HH140 on cultured keratinocytes (Table 2) There was a consistent tendency for a, b and g to increase with passage number. This increase was visible in the cell number, the DNA and the protein numbers. The values for all of these parameters were similar to those for the normalized cells subjected to HH16 (Table 1).
3.5. The effect of KTM HH76 on cultured keratinocytes (Table 3) The results for this drug recipe were very similar to those for HH140 (Table 2), except that the tendency to stimulate cell growth was smaller. For these cells, the drug, HH76, seemed to be quite
4. Discussion Our GC measurement results show no change in the membrane fatty acid composition following any of the drug additions to the media. As these drug preparations were aqueous solutions of herbs, no direct influence on the cellular lipids was expected. However, these drugs are used in the treatment of inflammatory and other chronic diseases so that their mode of action may be to change the inflammatory response by changing the cellular concentration of the polyunsaturated fatty acids, especially arachidonic acid. Because arachidonic acid is the main precursor for the lipoxygenase and the cyclooxygenase pathways, it is important to eliminate the possibility that these drugs work by increasing the concentration of this fatty acid or its precursors. Any change in EFA content could originate from modifications to the control of fatty acid synthesis or metabolism via genetic mechanisms. These pathways are not present in keratinocyte cultures; de novo synthesis of fatty acids from acetate ends in the formation of monounsaturated fatty acids. In normal tissue and in EFA-supplemented keratinocytes, there is no active metabolic pathway from 18:2 to 20:4 fatty acid. However, in EFA-deficient epidermal cells, the conversion of 18:2 to 20:4 fatty acid is extremely rapid (Marcelo and Dunham, 1993). Therefore, the genetic code to form arachidonic acid (from linoleic acid) is present in the nucleus but apparently not activated by these drugs. Cell growth as measured by cell count, DNA or protein levels gave consistent results for all the drugs studied. In all experiments, the data were analyzed as a percentage of the growth of the control culture for the experiment. This means that a value of 100 can be interpreted as a null effect on cell growth by the drugs, but it also means that the keratinocytes were growing very rapidly as is normal for the EFA-deficient keratinocyte cultures.
60
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
In Tables 1–3, the growth data for the three drugs are re-parameterized as three numbers: a, b and g. The mathematical model in Eq. (Eq. (1)) was chosen so that these three numbers are approximately equal if all three effects are similar in their effect on the growth estimator: cell number, DNA or protein values. a and b signify a stimulatory and regulatory effect, respectively, on cell growth rate; whereas g signifies a toxic effect on the cell culture. It is important to re-emphasize that both a and b can be large and the cell culture can still be rapidly growing and healthy. If g is large, then the cell culture is dead at high drug doses. Therefore, the two parameters, b and g, although seeming to be similar, have very different mathematical roles and represent very different biological consequences. One of the most satisfying aspects of the model in Eq. (Eq. (1)) is that, in the data in all three Tables, the value of a is correlated with the value of b. Statistically, half of the change in b can be attributed to a change in a. This correlation falls short of the 0.9 to 0.95 value that would be needed to imply a direct connection, but a close inspection of the data indicate that the lack of correlation is probably due to random errors in the measurements. Therefore, it is our judgment that Eq. (Eq. (1)) is probably a valid descriptor of the data in these experiments. The importance of this model is that it was obtained from our data, and it implies that the stimulatory effect of these drugs on cell growth was countered by a controlling influence that also came from the medicine. This Yang/Yin aspect to the data is an attractive surprise to our findings in that it is compatible with a major philosophical point of view of oriental medicine. Our results are easily summarized by recalling that only in the case of the KTM HH16 was the drug non-toxic to the cell system. For the high proliferative keratinocytes, the stimulatory/regulatory effect on cell growth was seen in the absence of any toxic effect on the cell strains (Table 1). When the cell strains had normal growth rates, the toxic effects of the drug addition to the medium masked this stimulatory effect. However, the stimulatory/regulatory effect was measurable and significant for all cell strains and drug combinations.
One main difference between the high proliferative cells (Table 1) and the EFA-supplemented cells (Table 1) is that the EFA-normalized cells are in a more advanced stage of differentiation than the EFA-deficient cells. It is possible therefore that HH16 displays its toxicity by interfering with some differentiation pathway that is present in the ‘normalized’ cell strains. This conclusion is further substantiated by the observation that, for all the drugs, the stimulatory/regulatory effect and the toxicity effect are seen to increase with passage number. The results for KTM HH140 and HH76 on proliferating EFA-deficient cells are similar to those for HH16 on normalized EFA-supplemented cells: a stimulatory/regulatory effect plus toxicity from the KTMs. It is difficult to conclude any unique importance to these results because the KTMs used had such divergent targets and herbal compositions. The compositions of KTMs have been developed over centuries by trial and error procedures that lead to the inclusion of the many ingredients in each KTM. The complication of these recipes is attributable in part by the desire to control side effects of some of the necessary ingredients in the recipes. Therefore, many of the most effective KTMs have elaborate recipes. Our choice was to use very effective recipes on keratinocytes to discover whether we could measure variation in a chosen set of biological parameters. We were successful in measuring these variations due to the KTM additions to the medium, but our results imply that this method of investigation is not likely to be successful in identifying a particular pathway for the mechanism of the drugs. The problem with our approach is that the mechanism of these drugs is too complicated (involving stimulation, regulation and toxicity) to be traceable by observing previously chosen cell parameters. Rather than use the approach followed by our experiments, we suggest the following alternative. Choose a traditional medicine whose effect is well understood and well defined (e.g., ‘increases T-cell count’ as opposed to ‘strengthens the immune system’). Instead of trying to discover the mechanism of an unknown drug system, one should start with a known physical effect of the drug and
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
trace backward to discover the mechanism for the drug. This method has the advantage of not depending on understanding the interactions of the drug components to enable an understanding of the drug mechanism. By focusing on the effect instead of the cause, one is guaranteed to be measuring at least one important physiological parameter at all times during the investigation. In spite of these shortcomings in our investigations, these experiments represent an important first step in the attempt to understand the mechanisms of KTMs in inflammatory processes on a cellular level. We have discovered that the drugs have stimulatory/regulatory effects on cell strains as well as toxic effects. Our studies with EFA-supplemented cell strains suggest that the KTMs used to treat inflammatory disease may have the capability to interact with cellular differentiation pathways of human keratinocytes. These investigations indicate that these medicines have complicated compositions presenting a challenge that involves identifying real effects of the KTM extracts as well as the synergistic relationships of the various components in the mixture.
Acknowledgements This work was supported by the Department of Surgery, Division of Plastic and Reconstructive Surgery, and the Biophysics Research Division, University of Michigan, MI, USA, and by the Kyung Hee University Medical Center, Seoul, Korea and by the U.S.P.H.S. via NIH grants AR26009 (Marcelo) and AR42223 (Dunham).
References Boyce, S.T., Ham, R.G., 1983. Calcium-regulated differentiation of nonnal human epidermal keratinocytes in chemically defined clonal culture and serum-free serial culture. Journal of Investigative Dermatology 81, 33–40. Boyce, S.T., Ham, R.G., 1985a. Cultivation, frozen storage, and clonal growth of non-nal epidermal keratinocytes in serum-free medium. Journal of Tissue Culture Methods 9, 83– 93.
.
61
Boyce, S.T., Ham, R.G., 1985b. Normal human epiden-nal keratinocytes. In: Weber, M.M., Sekely, L. (Eds.), In Vitro Models for Cancer Research. CRC Press, Boca Raton, FL, pp. 245 – 274. Burton, K., 1968. Determination of DNA concentration with diphenylamine. In: Grossman, L., Moldave, K. (Eds.), Methods in Enzymology, vol. XII. Academic Press, New York, pp. 163 – 166. Cohen, L.V., Garner, W.L., Marcelo, C.L., 1995. In vitro nutrient fatty acid effects on keratin expression and skin graft preparation. National Meeting of the Society for Investigative Dennatology, Journal of Investigative Dermatology, May 26. Doo, H. 1997. Kyung Hee Herb Formula Index’ (Korean). Oriental Medicine Hospital, Kyung Hee Medical Center, Seoul Korea, pp. 141, 335, 429. Falanga, V., Margolis, D., Alvarez, O., Auletta, M., Maggiacomo, F., Altman, M., Jensen, J., Sabolinski, M., HardinYoung, J., 1998. Rapid healing of venous ulcers and lack of clinical rejection with an allogeneic cultured human skin equivalent. Archives of Dermatology 134, 293 – 300. Hur, J. 1989. ‘Tong-eui-po-kam’ (Korean). Nam-san-dang, Seoul, Korea, ,pp. 372, 378,447. Langdon, R., Cuono, C., Madri, J., Birchall, N., 1988. Reconstitution of structure and cell function in human skin grafts derived from cryopreserved allogeneic dermis and autologous cultured keratinocytes. Journal of Investigative Dermatology 91, 478 – 485. Lowry, O.H., Rosenbrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry 193, 265 – 275. Marcelo, C.L., Duell, E.A., Rhodes, L.M., Dunham, W.R., 1992. An in vitro model of essential fatty aid deficiency. Journal of Investigative Dermatology 99, 703 – 708. Marcelo, C.L., Dunham, W.R., 1993. Fatty acid metabolism studies of human epidermal cell cultures. Journal of Lipid Research 34, 2077 – 2090. Marcelo, C.L., Rhodes, L.M., Dunham, W.R., 1994. Nornialization of essential-fatty-acid-deficient keratinocytes requires palmitic acid. Journal of Investigative Dermatology 103, 564568. Pittelkow, M.R., Scott, R.R., 1986. New techniques for the in vitro culture of human skin keratinocytes and perspectives on their use for grafting of patients with extensive burns. Mayo Clinic Proceedings 61, 771 – 777. Willie, J.J., Pittelkow, M.R., Shipley, G.D., Scott, R.E., 1984. Integrated control of growth and differentiation of normal human prokeratinocytes cultured in serum-free medium: clonal analyses, growth kinetics and cell cycle studies. Journal of Cell Physiology 121, 31 – 44. Wright, N.A., 1983. The cell proliferation kinetics of the epidermis. In: Goldsmith, L.A. (Ed.), Biochemistry and Physiology of the Skin. Oxford Press, New York, pp. 203 – 240.
The effect of three Korean traditional medicines on the growth rate of cultured human keratinocytes Seok Hee Chung a,1, Hiroto Terashi a, Lenore M. Rhodes a, Namdoo Moon b, William R. Dunham b,*, Cynthia L. Marcelo a a
Di6ision of Plastic and Reconstructi6e Surgery, Department of Surgery, Uni6ersity of Michigan Medical Centers, Ann Arbor, MI 48109, USA b Biophysics Research Di6ision, The Uni6ersity of Michigan, 930 North Uni6ersity A6enue, Ann Arbor, MI 48109 -1055, USA Received 1 July 2000; received in revised form 14 August 2000; accepted 21 August 2000
Abstract The effect of three different Korean Traditional Medicines (KTM) was studied on several functional parameters of adult human cells in culture. The cells were non-transformed strains of normal, skin epidermal cells (keratinocytes) from adult humans. Aqueous extracts of the herbal medicines were tested using two types of cell strains: one type was essential fatty acid deficient (EFAD) cells which grow rapidly in medium that was low in calcium and had no essential fatty acids; the second type was a cell strain grown in medium supplemented with essential fatty acid (EFA-supplemented). These cells had much slower, in vivo skin growth rates, and the fatty acid composition resembled that measured in epidermal biopsy tissue. The KTMs chosen for this study were tae-gang-hual-tang (for treating osteoarthritis), hual-ak-tang (for pain relief) and sip-zeon-tae-bo-tang (for fortifying immune systems). Because high proliferation rates usually correlate with skin inflammation and because many of the chemotactic agents mediating inflammatory response are modified fatty acids, this study focused on cell growth rate and membrane fatty acid composition as signals for the effects of the herbal medicines. By monitoring growth rate, these experiments measured both a stimulatory and a regulatory effect on the growth of keratinocytes. Some toxicity was seen at the highest doses of the KTMs. These effects were modeled mathematically, and the results showed varying effects on growth rate depending on dose and herbal recipe. The fitting parameters were discussed as they relate to biological function. The experimental design was also discussed and alternatives were suggested. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Epidermal cells; Epidermis; Gas chromatography; Alternative medicine; Mathematical model; Adult human cell strains
* Corresponding author. Tel.: + 1-734-7636721; fax: + 1-734-7643323. E-mail address: [email protected] (W.R. Dunham). 1 Present address: Department of Physiotherapy, College of Oriental Medicine, Kyung Hee University, Seoul 130-702, South Korea. 0378-8741/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 8 7 4 1 ( 0 0 ) 0 0 3 4 2 - 1
54
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
1. Introduction The medicines used in Korean Traditional Medicine (KTM) consist of appropriate amounts of mixed natural products, such as herbs, animal derivatives, and minerals. Each component is measured according to the principles of KTM theory, which considers synergistic effects and drug interactions. Typical KTMs have five to fifteen components, which interact to mollify any toxicity, to minimize side effects and to maximize the effect of the drug. In most cases, KTMs are mixed with water, boiled and extracted yielding a tea, which is drunk by the patient (Hur, 1989; Doo, 1997). One possible approach to studying these medicines is to identify their active ingredients and to test each one until a correlation is discovered between an ingredient and the curative effect of the medicine. We rejected this method of study for several reasons. Experience has shown that the three KTMs are valid healing agents and that all of the herbs and active ingredients are necessary for the medicine to be effective. Therefore, it is not possible to study only the effect of a component of a specific KTM; the entire recipe must be kept intact. Secondly, drug companies long ago have mined oriental medicines for their ingredients. Lists of the known active ingredients of these herbs are available in the references to this manuscript and elsewhere. These lists are long; to study each would not be possible in the time frame for this study. Furthermore, it is probable that this type of study would reproduce work already done in drug companies and, in view of the need to retain all of the KTM ingredients; we felt that this type of study would have been unlikely to produce new and useful information. Oriental medicines are effective treatments in most diseases and are most effective in the treatment of functional (no visible anatomical effects) or chronic diseases. Standard practice of KTM prescribes that mechanical or acute diseases are treated more effectively by the modern technology of Western medicine. In Korea, KTM treatment is most prevalent in the management of chronic illnesses in the elderly, which comprise a large portion of the ailing population. Even though
widely practiced, a fundamental, scientific understanding of the biochemical mechanism of any often-used KTM formulation is not available to the medical community. To simplify and increase the efficacy of KTM, it would be beneficial to develop standard methods to identify the main classes of active drugs in KTM. Thus far, all KTM drug studies have been performed using animal models. We therefore decided to develop methods to test the active drugs of KTM using a primary human epidermal cell culture system. It was decided to use this cell system because it is a cell strain whose life span has not been transformed by modification of its DNA. Accordingly, the cell strain has a limited life span, and these cells are grown under very well-defined medium conditions. The adult human epidermal cells are routinely used to create both temporary and permanent skin grafts for wound healing, forming ‘normal-appearing’ tissue (Pittelkow and Scott, 1986; Langdon et al., 1988; Falanga et al., 1998). The epidermal cells are harvested using the enzyme, trypsin, from split thickness skin that is routinely discarded from elective surgical procedure at the University of Michigan Medical Center. The resulting primary basal epidermal cell cultures are keratinocyte cultures. Growth of keratinocytes in serum- and lipid-free, low calcium conditions results in rapidly growing cultures that can be greatly expanded by numerous subculturing procedures (passages), without the loss of viability (Boyce and Ham, 1985a,b). The cultures grow well to the sixth passage when senescent (non-dividing) cells become the dominant cell type in the culture (Marcelo et al., 1992). Keratinocytes grown under these conditions are severely deficient in essential fatty acids (EFA) by the first passage (Marcelo et al., 1992). These cells have greatly diminished amounts of the major n-6 polyunsaturated fatty acids (18:2-, 20:3- and 20:4fatty acids), and contain increased amounts of the monounsaturated fatty acids, 16:1(n-7) and 18:1(n-9). Using published methods to control the fatty acid composition of normal adult human keratinocytes in culture (Marcelo et al., 1994), the fatty acid composition of the keratinocyte cultures were normalized to the levels found in human epider-
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
mal cell membranes. The fatty acid restored cultures presented a slowed growth rate, a decrease in successful passage and a more visually differentiated (flat and cornified) cellular appearance (Marcelo et al., 1992). Analysis of keratin type and cell envelope formation using PAGE and monoclonal antibodies led to the conclusion that the cells senesced more rapidly, rather than differentiated, under the influence of the EFA family of lipids (Cohen et al., 1995). This manuscript describes a study of the effects of three different KTMs on the growth rate and other physiologic parameters of human keratinocytes grown in both the EFA-deficient and -supplemented media. The KTMs were selected from 320 herb extract formulas commonly used at the Kyung Hee University Medical Center, Seoul, Korea: Tae-gang-hual-tang (HH 16) (Hur, 1989; Doo, 1997) is used for osteoarthritis, hual-ak-tang (HH 140) Hur, 1989 for low back pain, shoulder pain, and sciatica, and sip-zeon-tae-bo-tang (HH 76) (Hur, 1989; Doo, 1997) for fortifying the immune system. All of these medicines are expected to change the inflammatory response of the affected tissue. Because there is a large body of evidence concerning the effects of various reagents on skin growth, it was decided to use human epidermal keratinocytes as a platform for this study. The focus of our laboratory is the involvement of fatty acids in this response. Combining the interest in fatty acids with our experience with this keratinocyte culture technique led to the design of the experiments. For example, it has been found that cell proliferation and cell growth rate are often correlated in the human skin disease, psoriasis (Wright, 1983). Accordingly, the effect of these medicines on keratinocyte number, total protein and DNA, and cell fatty acid composition was investigated as gauge of the effect of these three KTMs on human tissue in general.
2. Material and methods
2.1. Keratinocyte culture The basic medium, MCDB 153, was prepared as previously described (Boyce and Ham, 1983,
55
1985a,b) and was supplemented with 0.6 mM (0.218 mg/ml) hydrocortisone, 5 ng/ml epidermal growth factor (EGF), 5 mg/ml insulin, 6 mg% bovine pituitary extract (3ml of extract, 10 mg protein/ml, in 500ml of medium) and 0.15 mM CaCl2 to form complete growth medium. All components of the medium, including the bovine pituitary extract, are lipid-free. Discarded human skin from mastectomy and abdominoplasty cosmetic surgery was rinsed and soaked for 2 h in solution A (30 mM hydroxyethypiprazine-N%-2-ethansulfonic acid (HEPES), 10 mM glucose, 3 mM KCl, 130mM NaCl, 1.0 mM Na2HPO4, pH 7.4) (Boyce and Ham, 1983). The skin was cut into 0.2–0.3 × 7 cm strips and digested overnight at room temperature with 0.08% trypsin in solution A, which separates the skin at the dermal-epidermal junction, leaving the basal cells on the dermal layer. The basal epidermal cells (keratinocyte) were gently scraped from the separated dermis into MCDB 153 plus 10% PBS. The filtered cell suspension (250 mm mesh, PGC Scientific, Gaithersburg, MD) was plated at 20 million cells in 15 ml MCDB 153 containing 2% chelexed FBS per T-75 flask and was grown at 37°C in 5% CO2 in air gassing. The cultures were fed every 48 h with serum-free MCDB 153 and reached approx. 70% confluence after 3–5 days (Boyce and Ham, 1983, 1985a,b; Willie et al., 1984). Usually, the monolayers were passaged every 3rd to 5th day, yielding an expanding strain from each surgical specimen. 2.2. Fatty acid supplementation The stock fatty acids (18:2 (n-6), 16:0 and 20:4 (n-6)) were stored in chloroform (Marcelo et al., 1994). A 30× final concentration (10 mM 18:2 and 5 mM 16:0 and 20:4 fatty acids) was evaporated under an N2 stream (sterilized using a Millex–SLGS, Millipore filter Millipore, Boston, MA). Ten microliters of 95% ethanol was added to the tube to solubilize the fatty acid, plus 1/30th final volume of sterile 0.6% bovine serum album in medium. After vigorous vortexing, the fatty acids plus carrier were added to the medium. The final amount of bovine serum albumin was 0.02%. The medium was freshly prepared at each change of medium.
56
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
2.3. Herb preparation and extract supplementation KTMs were prepared by the Kyung Hee University oriental medical hospital. Each prescription is shown in the following section. All herbs were dried and weighed. The prescription was added to a 10× weight of water, boiled for 1.5 h, and extracted. The extracted tea was then filtered using a centrifuge, concentrated by vacuum falling filter evaporator, dried by spray dryer, and granulated (Doo, 1997). Specifically, 6 g of herb extract was dissolved in medium to a final concentration of 0.6 g/10 ml for 1 h at 37°C. It was then filtered-sterilized and added to the MCDB 153 medium. The keratinocytes were fed with this medium every 48 h. Cells were grown in both the control (EFA-deficient) and fatty acid supplemented (EFA-supplemented) media.
2.4. Herb formulae 2.4.1. HH16(Tae-gang-hual-tang) Ingredients by descending order of weight: Notopterygium incisum, Cimicifuga heracleifolia, Angelica pubescens f. biserrata (6 g each); Akebia quinata, Chaenomeles sinensis, Achyranthes bidentata, Areca catechu (4 g each); Atractylodes japonica, Atractlodes macrocephala, Stephania tetrandra, Clematis manshurica, Angelica gigas, Poria cocos, Crataegus pinnatifida var. major, Hordeum 6ulgare var. hexastichon, Alisma plantago-aquatica var. orientale, Glycyrrhiza uralensis (3 g each). 2.4.2. HH140 (Hual-ak-tang) Ingredients by descending order of weight: Chaenomeles sinensis, Chelidonium majus (10 g each), Corydalis turtschaninovi (8 g), Clematis manshurica, Angelica pubescens f. biserrata, Angelica gigas, Rehmannia glutinosa, Paeonia lactiflora, Atractylodes japonica (6 g each); Citrus unshiu, Boswellia carterii, Commiphora myrrha, Zingiber officinale, Zizyphus jujuba var. inermis (4 g each); Carthamus tinctorius, Amomum villosum (3 g each); Glycyrrhiza uralensis (2 g).
2.4.3. HH 76 (Sip-zeon-tae-bo-tang) Ingredients by descending order of weight: Panax ginseng, Atractylodes macrocephala, Poria cocos, Rehmannia glutinosa, Paeonia lactiflora, Cnidium officinale, Angelica gigas (6 g each); Glycyrrhiza uralensis, Astragalus membranaceus, Cinnamomum cassia (4 g each). 2.5. Analysis of fatty acid methyl esters For lipids analysis, the cell monolayers were drained of medium, rinsed twice with cold phosphate buffered saline (PBS). Two milliliters of HPLC grade methanol was added over the cell layer. After scraping the monolayer, the material was then extracted with a 1:2:1.5 ratio of methanol:chloroform:0.1M KCl in 50% methanol, and the organic phase was processed as previously reported (Marcelo et al., 1992). The sample was evaporated under nitrogen stream, re-suspended in 50 l of 1:1 chloroform:methanol, applied to a thin-layer chromatography (TLC) silica gel 60 plate (Merck, Darmstadt, Germany) and chromatographed in one direction using CHCl3:MeOH:acetic acid (90:8:1). All the phospholipids, which remained at the origin, were scraped and the lipids eluted from the silica during transmethylation with 6% methanolic–HCl+ 50 mg 17:0 as an internal, quantitative standard. The fatty acid methyl esters were extracted into petroleum ether (Baker, Phillipsburg, NJ), evaporated to dryness and stored frozen under benzene. For analysis, the samples were evaporated to 200 ml of benzene, filtered using a 0.45 mm filter, were evaporated and re-suspended in 100–150 ml filtered chloroform. One microliter chloroform was injected into the gas chromatograph for analysis. The samples were analyzed using a Hewlett Packard (Wilmington, DE) GC Model 5710A equipped with a J & W Scientific (Folsom, CA) fused silica Megabore DB225, 0.53-m diameter column. A 21-fatty acid standard (Nu-CHEKPREP, Inc., Elysian, MN) was used to calibrate the GC elution times and to identify the fatty acids. A known amount of 17:0-fatty acid was added as an internal standard, because its peak is easily identifiable and quantified when the chro-
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
57
matogram is digitized by an IBM-PC computer interface (Model AN-146; Alpha Products, Darien, CT). We wrote the recording and data evaluation software in BASIC or FORTRAN. The data are presented as mg% or as mg fatty acid per mg of protein.
T-distribution (standard techniques). The parameter values in the tables are the means (N= 3).
2.6. Cell proliferation studies
3.1. Fatty acid composition
For studying cell proliferation, the cultures were lightly trypsinized and counted using a hemocytometer; trypan blue exclusion was used to determine cell viability. DNA Burton, (1968) and protein (Lowry et al., 1951) assays were used to quantitate these materials. The control cells (no KTM) were allowed to grow to 70-80% confluence (3-5 days); then both the control- and KTMtreated cell numbers were determined.
Fatty acid composition of the cellular lipids was monitored by gas chromatography (GC). For each passage, the lipids of the control and KTMtreated cells were measured and compared. In no case was there a change in the fatty acid composition due to the addition of the KTMs to the cell growth medium. Our typical results are shown in Fig. 1, where the GC chromatogram of P3 (passage three) EFA-deficient control cells is plotted. Compared to these control cells, the EFA-supplemented cells presented a greater 18:2-fatty acid peak (7.8 min), a smaller 16:1-fatty acid peak (4.5 min), a smaller 18:1-fatty acid (7.1 min) peak and much greater 20:3- and 20:4-peaks (13.9 and 14.4 min).
2.7. Statistical analysis Each experiment was performed three times, each with a different cell strain (from a different primary culture). As we will show below, the data were analyzed using curve-fitting procedures to yield three parameters for each experiment: alpha, beta and gamma. These values (N =3) were analyzed to provide information on statistical significance of the values in the tables using the
Fig. 1. GC spectra of the fatty acids in EFA-deficient keratinocytes. The elution times are: 2.9 min= 14:0; 4.2 min= 16:0; 4.5 min= 16:1; 5.3 min =17:0 (50 mg internal standard); 6.7 min= 18:0; 7.1 min =18:1; 7.8 min = 18:2; 12.0 min = 20:1; 14.4 min =20:4; 18.3 min = 22:0; 27.5 min = 24:0; and 28.9 min= 24:1 fatty acid. The identifications are made using the calibration standard shown in the text.
3. Results
3.2. Cell growth rates Cell growth as a function of drug dose was determined by measuring the total number of cells in the cultures, the amount of DNA in the cultures, and the amount of protein in the cultures. All measurements were converted to ‘% control’ values by dividing the measurements by the appropriate control (no drug) and multiplying by 100. The data from one of the dose/response tests of the KTM (HH16) cells are shown in Fig. 2. These data are representative of all of the results for all the combinations of drugs and cells. At the lowest drug additions to the media (both EFAdeficient and supplemented), the growth numbers were larger than 100, indicating an enhancement of the growth numbers compared to the control values. At higher dose values, this enhancement effect decreased with increasing dose values. In some cases, the cell population as well as the DNA and protein values decreased to a value far smaller than 100, which indicated that the drug was toxic to the cell culture.
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
58
Fig. 2. Modeling of the Effects of KTM on keratinocytes in culture. Cell growth as a function of drug dose was determined by measuring the total number of cells in the cultures, expressed as ‘% control’. In this figure is presented the data from one of the dose/response tests of the KTM (HH16) cells. These data are representative of all of the results for all the combinations of drugs and cells. Table 1 Cell number, DNA, and protein fit parameters [Eq. (Eq. (1))] for the cells grown in EFA-deficient and supplemented medium and HH16 KTM. P I, P2 and P3 refer to consecutive passages of the cell strains, P I being the first passage after the primary culture of the keratinocytes a
P1 P2 P3 P1 P2 P3 P1 P2 P3 P1 P2 P3
b
EFA-deficient Cell number 0.041 0.015 0.047 0.018 0.042 0.029 EFA-deficient DNA 0.012 0.025 0.024 0.031 0.078 0.072 EFA-deficient Protein 0.010 0.022 0.030 0.011 0.038 0.030 EFA-supplemented Cell Number 0.014 0.015 0.026 0.026 0.062 0.040
g
0.004 0.004 0.004 0.001 0.001 0.001 0.001 0.011 0.001 0.005 0.018 0.038
We have mathematically modeled these two effects by the following equation, G = 100axe-bx + 100e-gx
(1)
where G signifies the cell growth (cell c , DNA or protein level); x signifies the dose added to the medium (ml); and a, b and g are data fitting parameters. Both terms in the above expression contain exponential functions, which result mathematically from the assumption that the attenuation of the relevant term is proportional to the dose. In the first term, the enhancement of cell growth, ax, is attenuated by a control argument, bx. If a is equal to zero, then the size of b is irrelevant because there is no enhancement to control. If the toxicity parameter, g, is equal to zero, then the value of G remains equal to 100 when the first term becomes negligible at very large doses. A larger value for g indicates a higher toxicity. In Table 1, Table 2, and Table 3 are listed measured values of the fitting parameters at passage 1, 2, and 3 for each of the drug compositions, HH16, HH140 and HH76. In all cases except for the HH16 drug, the cells were grown in medium with no fatty acid. For the HH16 drug tests, the cells were also grown in fatty acid enhanced medium as indicated in Table 1.
3.3. The effect of KTM HH16 on cultured keratinocytes (Table 1) For the cell cultures grown without fatty acids in the culture medium, EFA-deficient, g was not significantly different from zero; in other words, Table 2 Cell number, DNA, and protein fit parameters [Eq. (1)- for the cells grown inEFA-deficient medium and HH140 KTM a
P1 P2 P3
0.009 0.027 0.098
P1 P2 P3
0.009 0.027 0.079
P1 P2 P3
0.008 0.030 0.072
b Cell number 0.008 0.027 0.030 DNA 0.008 0.027 0.035 Protein 0.008 0.013 0.025
g
0.007 0.008 0.030 0.008 0.008 0.009 0.008 0.013 0.025
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61 Table 3 Cell number, DNA, and protein fit parameters [Eq. (Eq. (1))] for the HH 176 EFA deficient cells a
P1 P2 P3
0.014 0.059 0.041
P1 P2 P3
0.000 0.118 0.127
P1 P2 P3
0.041 0.037 0.058
b Cell number 0.000 0.026 0.033 DNA 0.108 0.058 0.048 Protein 0.033 0.016 0.024
59
toxic to the cell cultures at the highest passage number compared to the results from the test of the other two drug recipes, HH16 and HH140.
g
0.019 0.000 0.033 0.000 0.000 0.048 0.008 0.000 0.023
there was no significant toxicity from KTM HH16 for these cells. For the EFA-deficient (highly proliferative) cells, the value of a (cell number) was constant at an average value of 0.044 90.2 and the value of b was fairly constant at 0.0219 0.005 (average9 SEM); however, the DNA and protein values showed a tendency to increase with passage number. This tendency was also seen in the EFA-supplemented (normal growth rate and fatty acid composition) cells, where the value of g also showed this tendency to increase with passage number.
3.4. The effect of KTM HH140 on cultured keratinocytes (Table 2) There was a consistent tendency for a, b and g to increase with passage number. This increase was visible in the cell number, the DNA and the protein numbers. The values for all of these parameters were similar to those for the normalized cells subjected to HH16 (Table 1).
3.5. The effect of KTM HH76 on cultured keratinocytes (Table 3) The results for this drug recipe were very similar to those for HH140 (Table 2), except that the tendency to stimulate cell growth was smaller. For these cells, the drug, HH76, seemed to be quite
4. Discussion Our GC measurement results show no change in the membrane fatty acid composition following any of the drug additions to the media. As these drug preparations were aqueous solutions of herbs, no direct influence on the cellular lipids was expected. However, these drugs are used in the treatment of inflammatory and other chronic diseases so that their mode of action may be to change the inflammatory response by changing the cellular concentration of the polyunsaturated fatty acids, especially arachidonic acid. Because arachidonic acid is the main precursor for the lipoxygenase and the cyclooxygenase pathways, it is important to eliminate the possibility that these drugs work by increasing the concentration of this fatty acid or its precursors. Any change in EFA content could originate from modifications to the control of fatty acid synthesis or metabolism via genetic mechanisms. These pathways are not present in keratinocyte cultures; de novo synthesis of fatty acids from acetate ends in the formation of monounsaturated fatty acids. In normal tissue and in EFA-supplemented keratinocytes, there is no active metabolic pathway from 18:2 to 20:4 fatty acid. However, in EFA-deficient epidermal cells, the conversion of 18:2 to 20:4 fatty acid is extremely rapid (Marcelo and Dunham, 1993). Therefore, the genetic code to form arachidonic acid (from linoleic acid) is present in the nucleus but apparently not activated by these drugs. Cell growth as measured by cell count, DNA or protein levels gave consistent results for all the drugs studied. In all experiments, the data were analyzed as a percentage of the growth of the control culture for the experiment. This means that a value of 100 can be interpreted as a null effect on cell growth by the drugs, but it also means that the keratinocytes were growing very rapidly as is normal for the EFA-deficient keratinocyte cultures.
60
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
In Tables 1–3, the growth data for the three drugs are re-parameterized as three numbers: a, b and g. The mathematical model in Eq. (Eq. (1)) was chosen so that these three numbers are approximately equal if all three effects are similar in their effect on the growth estimator: cell number, DNA or protein values. a and b signify a stimulatory and regulatory effect, respectively, on cell growth rate; whereas g signifies a toxic effect on the cell culture. It is important to re-emphasize that both a and b can be large and the cell culture can still be rapidly growing and healthy. If g is large, then the cell culture is dead at high drug doses. Therefore, the two parameters, b and g, although seeming to be similar, have very different mathematical roles and represent very different biological consequences. One of the most satisfying aspects of the model in Eq. (Eq. (1)) is that, in the data in all three Tables, the value of a is correlated with the value of b. Statistically, half of the change in b can be attributed to a change in a. This correlation falls short of the 0.9 to 0.95 value that would be needed to imply a direct connection, but a close inspection of the data indicate that the lack of correlation is probably due to random errors in the measurements. Therefore, it is our judgment that Eq. (Eq. (1)) is probably a valid descriptor of the data in these experiments. The importance of this model is that it was obtained from our data, and it implies that the stimulatory effect of these drugs on cell growth was countered by a controlling influence that also came from the medicine. This Yang/Yin aspect to the data is an attractive surprise to our findings in that it is compatible with a major philosophical point of view of oriental medicine. Our results are easily summarized by recalling that only in the case of the KTM HH16 was the drug non-toxic to the cell system. For the high proliferative keratinocytes, the stimulatory/regulatory effect on cell growth was seen in the absence of any toxic effect on the cell strains (Table 1). When the cell strains had normal growth rates, the toxic effects of the drug addition to the medium masked this stimulatory effect. However, the stimulatory/regulatory effect was measurable and significant for all cell strains and drug combinations.
One main difference between the high proliferative cells (Table 1) and the EFA-supplemented cells (Table 1) is that the EFA-normalized cells are in a more advanced stage of differentiation than the EFA-deficient cells. It is possible therefore that HH16 displays its toxicity by interfering with some differentiation pathway that is present in the ‘normalized’ cell strains. This conclusion is further substantiated by the observation that, for all the drugs, the stimulatory/regulatory effect and the toxicity effect are seen to increase with passage number. The results for KTM HH140 and HH76 on proliferating EFA-deficient cells are similar to those for HH16 on normalized EFA-supplemented cells: a stimulatory/regulatory effect plus toxicity from the KTMs. It is difficult to conclude any unique importance to these results because the KTMs used had such divergent targets and herbal compositions. The compositions of KTMs have been developed over centuries by trial and error procedures that lead to the inclusion of the many ingredients in each KTM. The complication of these recipes is attributable in part by the desire to control side effects of some of the necessary ingredients in the recipes. Therefore, many of the most effective KTMs have elaborate recipes. Our choice was to use very effective recipes on keratinocytes to discover whether we could measure variation in a chosen set of biological parameters. We were successful in measuring these variations due to the KTM additions to the medium, but our results imply that this method of investigation is not likely to be successful in identifying a particular pathway for the mechanism of the drugs. The problem with our approach is that the mechanism of these drugs is too complicated (involving stimulation, regulation and toxicity) to be traceable by observing previously chosen cell parameters. Rather than use the approach followed by our experiments, we suggest the following alternative. Choose a traditional medicine whose effect is well understood and well defined (e.g., ‘increases T-cell count’ as opposed to ‘strengthens the immune system’). Instead of trying to discover the mechanism of an unknown drug system, one should start with a known physical effect of the drug and
S. Hee Chung et al. / Journal of Ethnopharmacology 74 (2001) 53–61
trace backward to discover the mechanism for the drug. This method has the advantage of not depending on understanding the interactions of the drug components to enable an understanding of the drug mechanism. By focusing on the effect instead of the cause, one is guaranteed to be measuring at least one important physiological parameter at all times during the investigation. In spite of these shortcomings in our investigations, these experiments represent an important first step in the attempt to understand the mechanisms of KTMs in inflammatory processes on a cellular level. We have discovered that the drugs have stimulatory/regulatory effects on cell strains as well as toxic effects. Our studies with EFA-supplemented cell strains suggest that the KTMs used to treat inflammatory disease may have the capability to interact with cellular differentiation pathways of human keratinocytes. These investigations indicate that these medicines have complicated compositions presenting a challenge that involves identifying real effects of the KTM extracts as well as the synergistic relationships of the various components in the mixture.
Acknowledgements This work was supported by the Department of Surgery, Division of Plastic and Reconstructive Surgery, and the Biophysics Research Division, University of Michigan, MI, USA, and by the Kyung Hee University Medical Center, Seoul, Korea and by the U.S.P.H.S. via NIH grants AR26009 (Marcelo) and AR42223 (Dunham).
References Boyce, S.T., Ham, R.G., 1983. Calcium-regulated differentiation of nonnal human epidermal keratinocytes in chemically defined clonal culture and serum-free serial culture. Journal of Investigative Dermatology 81, 33–40. Boyce, S.T., Ham, R.G., 1985a. Cultivation, frozen storage, and clonal growth of non-nal epidermal keratinocytes in serum-free medium. Journal of Tissue Culture Methods 9, 83– 93.
.
61
Boyce, S.T., Ham, R.G., 1985b. Normal human epiden-nal keratinocytes. In: Weber, M.M., Sekely, L. (Eds.), In Vitro Models for Cancer Research. CRC Press, Boca Raton, FL, pp. 245 – 274. Burton, K., 1968. Determination of DNA concentration with diphenylamine. In: Grossman, L., Moldave, K. (Eds.), Methods in Enzymology, vol. XII. Academic Press, New York, pp. 163 – 166. Cohen, L.V., Garner, W.L., Marcelo, C.L., 1995. In vitro nutrient fatty acid effects on keratin expression and skin graft preparation. National Meeting of the Society for Investigative Dennatology, Journal of Investigative Dermatology, May 26. Doo, H. 1997. Kyung Hee Herb Formula Index’ (Korean). Oriental Medicine Hospital, Kyung Hee Medical Center, Seoul Korea, pp. 141, 335, 429. Falanga, V., Margolis, D., Alvarez, O., Auletta, M., Maggiacomo, F., Altman, M., Jensen, J., Sabolinski, M., HardinYoung, J., 1998. Rapid healing of venous ulcers and lack of clinical rejection with an allogeneic cultured human skin equivalent. Archives of Dermatology 134, 293 – 300. Hur, J. 1989. ‘Tong-eui-po-kam’ (Korean). Nam-san-dang, Seoul, Korea, ,pp. 372, 378,447. Langdon, R., Cuono, C., Madri, J., Birchall, N., 1988. Reconstitution of structure and cell function in human skin grafts derived from cryopreserved allogeneic dermis and autologous cultured keratinocytes. Journal of Investigative Dermatology 91, 478 – 485. Lowry, O.H., Rosenbrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry 193, 265 – 275. Marcelo, C.L., Duell, E.A., Rhodes, L.M., Dunham, W.R., 1992. An in vitro model of essential fatty aid deficiency. Journal of Investigative Dermatology 99, 703 – 708. Marcelo, C.L., Dunham, W.R., 1993. Fatty acid metabolism studies of human epidermal cell cultures. Journal of Lipid Research 34, 2077 – 2090. Marcelo, C.L., Rhodes, L.M., Dunham, W.R., 1994. Nornialization of essential-fatty-acid-deficient keratinocytes requires palmitic acid. Journal of Investigative Dermatology 103, 564568. Pittelkow, M.R., Scott, R.R., 1986. New techniques for the in vitro culture of human skin keratinocytes and perspectives on their use for grafting of patients with extensive burns. Mayo Clinic Proceedings 61, 771 – 777. Willie, J.J., Pittelkow, M.R., Shipley, G.D., Scott, R.E., 1984. Integrated control of growth and differentiation of normal human prokeratinocytes cultured in serum-free medium: clonal analyses, growth kinetics and cell cycle studies. Journal of Cell Physiology 121, 31 – 44. Wright, N.A., 1983. The cell proliferation kinetics of the epidermis. In: Goldsmith, L.A. (Ed.), Biochemistry and Physiology of the Skin. Oxford Press, New York, pp. 203 – 240.