Alteration of Proline Hydroxylase Activity by Corticosteroids

Cutroneo, Kenneth Robert. Ph.D., University of Rhode Island, September, 1970. Alteration of Proline Hydroxylase Activity by Corticosteroids. Major Professor: Dr. George C. Fuller. iii The biochemical mechanism by which glucocorticoids affect collagen metabolism was investigated . Triamcinolone, hydrocortisone and methylprednisolone significantly decreased liver proline hydroxylase activity in vivo. This was not a manifestation of an antianabolic effect on protein. Triamcinolone inhibited liver proline hydroxylase activity in a dose dependent manner in vivo. An observed elevation of liver hydroxylase activity in adrenalectomized animals was decreased to control level by hydrocortisone treatment. Subdermal implants of three types of sponges (polyurethane , cellulose and polyvinyl) were used to induce granuloma growth as a model system of inflam.~ation. The anti-inflammatory activity of corticosteroids was correlated with inhibition of proline hydroxylase activity in granuloma tissue . . The data indicate that corticosteroids decrease collagen syntre sis by inhibiting the rate limiting enzyme. This inhibitory effect on collagen metabolism is correlated with observed anti-inflamma tory activity.

have been reported (Hench et al., 1949(Hench et al., , 1950. These workers showed that large doses of ACTH or cortisone administered to patients suffering from rheumatoid arthritis produced dramatic and prompt improvement. 1 Although much is known about various metabolic and biochemical changes which occur in connective tissue at intervals after glucocorticoid treatment, little is known about the molecular and biochemical mechanism of action of this group of frequently used drugs on connective tissue metabolism. The ability of this class of drugs to impair the progress of connective tissue diseases is believed to be primarily due to the anti-inflarrunatory effect of these steroids (Lorenzen, 1969a).
Increased synthesis of collagen occurs as an essential function of the inflarrunatory process (Gilman, 1968). The effect of glucocorticoids on collagen me tabolism is well docume nted. Collagen content (Houck and Patel, 1965) and collagen s y nthesis (Fukuhara and Tsuru f ugi, 1969) are decre a s e d a £ter cortico i d a dmini s tration.
This i nve s tiga tion w~s con d u cte d to el u c i d at e the b i ochemica l mechanism b y wh ich t he gl u cocor ti coids a ffe c t coll agen synthesis and t o re l ate t his ac t i on t o anti-inf l anuna t ory activ ity.

A. The Inflanunatory Process and Proposed Modes of Action of Anti-inflarrunatory Steroids
Inflanunation is the series of biological processes involved in the response of tissues to injury. The two phases of inflanunation are the destructive phase and the reparative phase (Houck, 1968 (Whitehouse, 1968). Increased vascular permeability results from proteolytic enzyme destruction of the endo~ thelial vasculature. Aggregation of plateletsand conversion of fibrinogen to fibrin by an uncharacterized enzyme unique to inflammed connective tissue (Houck et al., 1967) results in thrombus formation and subsequent ischemia of the wounded area. Anoxia and acidosis of the tissue in turn causes necrosis (Houck, 1968).
A drug could pro.duce anti-inflanunatory activity by affecting one or more of the biochemical events of the inflammatory process. Various workers have shown that natural glucocorticoids and synthetic steroids stabilize lysosomes (Weissmann,1965a & b), possible mediators of inflammation (Weissmann, 1967). Willoughby and Spector (1964) reported that anti-inflammatory drugs are capable of inhibiting both proteases and esterases which attack the microcirculation. Gladner and Houck (1969) showed cortisol mediated the release of at least three proteases, one of which is highly specific for the phenyl-alanineserine bond in bradykinin. Anti-inflammatory drugs have also been shown to suppress the formation of histamine, 5-hydroxytryptamine and possibly the kinins, all of which are increased during inflammation (Whitehouse, 1968). Mustard et al. (1967) showed that anti-inflammatory drugs inhibit the aggregation of platelets. and has a molecular weight of 300,000 (Hutton et al.,4 1968). Collagen is composed of approximately one-third glycine, one-third proline and hydroxyproline and one-third other nonsulfur-containing amino acids (Crick and Rich , 1957). It contains no tryptophane and c ystine (Udenfriend, 1966). Collagen is the only protein that contains significant amounts of hydroxyproline .
Collagen biosynthesis involves two processes: (1) polypeptide synthesis and (2) hydroxylation of peptide-bound proline (Juva , 1968). The first step is probably similar to the synthesis of other proteins but the second is un ique and might very well prove to be the s·i te of regulation in the colla gen biosynthetic pathway. Many workers feel that protocollagen proline hydroxylase is the rate limiting step in collagen synthesis (Udenfriend, 1966). 5 Protocollagen, a soluble protein, serves as the substrate for the hydroxylation step . When protocollagen molecules reach molecular weights considerably greater than 10,000 and contain specified amino acid sequences which are recognized by the hydroxylase enzyme, specific proline residues are hydroxylated . Bhatmagar et al. (1967) believe that hydroxylation of proline occurs after the completed polypeptides of protocollagen are released from the ribosomes. However, Gould (1968) reports that hydroxyproline is associated with polyribosomes. Hydroxylation of free proline does not occur.
The oxygen atom of the hydroxyl group of hydroxyproline is derived from molecular o xygen and not from water (Fujimoto & Tamiya, 1962;Prockop et al., 1963).
A direct-displacement hydroxylation by this enzyme was demonstrated by Fujita et al. (1964). These workers using cis and trans-4-monotritio-proline showed that only the tritium atom trans to the carboxy l group was -lost during hydroxylation~ Later it was shown by Gottlieb et al. (1965) that trans-4-fluroproline was incorporated into collagen and converted to hydroxyproline. Thus, the enzyme is classified as a mixed function oxidase. Oxygen and a reducing agent are requ ired cofactors.
The enzyme has been demonstrated in a number of animal tissues Takeuchi et al., 1967;Mussini et al., 1967;Fuller and Langner, 1970). The enzyme activity is greatest in tissues actively engaged in collagen synthesis such as 6 embryonic and granuloma tissue (Mussini et al., 1967) and uterine tissue (Halme & Jaaskelainen, 1970).
The enzyme has been partially purified from foetal rat skin (Rhoads & Udenfriend, 1968) and has been shown to have an absolute requirement for atmospheric oxygen, ascorbate, -ketoglutarate and the ferrous ion (Hutton et al., 1967). The importance of ascorbate was shown by Stone and Meister (1962) who reported that hydroxyproline synthesis was disproporti.onate to proline incorporation in scorbutic tissue. Hutton et al. (1967) suggested that -ketoglutarate is an allosteric activator of the enzyme and its cellular concentration regulates the activity of the enzyme and the synthesis of collagen. However, Rhoads and Udenfriend (1968) demonstrated a stoichiometric d~carboxylation of ~ ketoglutarate coupled to the hydroxylation of peptidyl proline. These workers feel that oC -ketoglutarate serves as a specific electron donor, while ascorbate maintains a prope r reducing environment, and the ferrous ion acts as an electron carrier.
Since the enzyme involved in the hydroxylation reaction requires oxygen and ferrous ion, unhydroxylated protocollagen peptides may be prepared by incubating embryonic or connective tissue in the absence of oxygen or a chelator of iron, e.g. c<..OC:::dipyrydyl (Prockop & Juva, 1965) .
The assay of Hutton et al . (1966) is a rapid and specific method to determine collagen proline hydroxylase activity. The method is based on the enzymatic conversion of L-3,4 3 -H-peptidyl proline to L-3-3 H-hydroxyproline and release of a tritium atom from the four position.
The released tritium then equilibriates with unlabeled water. Tritiated water is collected and counted. It was shown that equivalent amounts of L-3-3 H-hydroxyproline and tritiated water were formed.

C. Glucocorticoids and Collagen Metabolism
Glucocor ticoid mediated decrease in collagen content in norma l connective tissue (Houck et al., 1967) and inflamed tissue (Fukuhara & Tsurufuji, 1969;Bavetta et al., 1962) is well documented; however, the mechanism of action is not known. The decrease in collagen content observed after corticoid treatment in animal and man could be explained by a decrease in the rate of collagen synthesis or an increased rate of collagen degradation.
Houck et al. (1967) reported a decrease in cutaneou s insoluble collagen in cortisol treated rats. Paralleling this loss in insoluble collagen was the appearance of cutaneous extracellular, free collagenolytic activity. These , authors feel that the observed loss in collagen is the result of collagen catabolism through the induction of collagenase activity in skin fibroblasts (Houck & Sharma , 1969). Kivirikko et al. (1965)  These results suggest that corticoids decrease coll agen synthesis.
Furthermore, in a related study these same workers demonstrated that cortisone did not affect the catabolism of insoluble collagen, The insoluble collagen of rat skin is relatively inert _metabolically with a biological half-life of one year (Lindstedt & Prockop, 1966). It is reasonable to assume that the corticoids would change collagen content by altering collagen biosynthesis rather than altering collagen degradation. Further, if the corticoids degraded mature collagen 1 urinary hydroxyproline levels should have been elevated. Neither Kivirikko et al. (1965) nor Smith (1967) observed catabolic changes in urinary hydroxyproline analyses. Urinary hydroxyproline levels were decreased in both studies. Smith (1967)  These workers found that collagen co·ntent rapidly increased following the seventh day after implantation. Robertson and Schwartz (1953) reported both biochemical and histological evidence of collagen fonna tion in granuloma tissue. The gra nuloma system provides an excellent model to study chronic in vivo stimulation of the proliferation of fibroblasts, synthesis of collagen, and the biochemic a l events involved in the regulation of these processes. Compared to o ther methods of granulation tissue developmen t, e.g.
carrageen in, the sponge-induced granuloma has the distinct advantage of being homogeneous and c an be separated from the surrounding tissue (Juva, 1968).
Glucocorticoids are well known inhibitors of granuloma formation (Fukuhara & Tsurufuji, 1969); however, the biochemical mechanism of this antigranulomatic activity has not yet been clearly elucidated. These workers showed that betamethasone inhibited collagen synthesis in the carageenin induced inflammatory tissue. Juva (1968) showed that changes 11 in protocollagen proline hydroxylase activity paralleled the rate of collagen synthesis in sponge induced granuloma tissue.
Since various workers feel that proline hydroxy lase is the rate limiting step in collagen biosynthesis (Mussini et al., 1967;Takeuchi et al., 1967;Gribble et al., 1969), the regulation of this enzyme may be the mechanism by which glucocorticoids affect collagen synthesis and granuloma formation.

F. Protocollagen Proline Hydroxy lase Assay
Proline hydroxylase activity in the 15,000 xg supernatant was measured by the method of Hutton et al. (1966). The samples were nex t allowed to stand at room temperature to cool. A 0.3 ml aliquot of each sample was added to 1.0 ml of 0.5 N NaOH and 5.0 ml a of Reagent A , mixed and allowed to stand for 20 minutes. Next 1 0.5 ml of b Reagent B was added to each sample, mixed and allowed to stand at room temperature for 40 minutes for color development.
The absorbance at 500 mu was read on a Beckman DB spectrophotometer.

H. Statistical Methods
The Student's-t-Test (Snedecor, 1956) was used to test for differences between means throughout. this investigation.
The formula employed a~ the basis for the comput.er programs used is as follows: where: df taken as: where: xl -_ x2 t = s*j 1 Steroid induced changes of distinct metabolic activities i n connec tive tissue may resµlt from a general inhibition of prote in synthesis (Davidson, 1963) or from an unspecifi c depres sion of a basic metabolic process (Nocenti, et al., 1964). It is unlikely that the · inhibitory effect on proline hydroxylase activity is a Enzyme activity is exDressed as the mean (+ S.E.) of the amount (dpm) of (::SH)H20 formed from (3-;4-3H) proline per milligram of protein per 30 minutes.
Anima ls received 150 mg/kg of drug i.p. for four consecutive days and were sacrificed 18 hours after last injection.
Significantly different from control at P < .05.
manifestation of an anti-anabolic effect on supernatant protein since protein concentration could be further decreased when praline hydroxylase inhibition .by hydrocortisone was maximal (Table 2). This argument is further . streng thened by the observation that triamcinolone decreased praline hydroxylase activity by 70 percent in the absence of an effect on supernatant protein (Table 3).
Triamcinolone inhibition of liver praline hydroxylase activity is linearly related to the number of daily doses ( Figure 1). Significant inhibition of enzyme activity was observed 24 hours after drug treatment. This dose dependent relationship suggests that the hydroxylase enzyme is the biochemical target for glucocorticoid induced alteration of collagen metabolism. Smith (1967) administered cortisone for at least 4 days before any change in insoluble skin collagen was observed. Thus, the anti-anabolic effect is seen much earlier than a catabolic effect on the insoluble pool as proposed by Houck et al. (1967).
Liver praline hydroxylase activity was elevated in adrenalectomized animals and was decreased to the con trol level by hydrocorti sone tre atment for four consecutiv Enzyme activity is ex~ressed as the mean (+ S E.) of the amount (dpm) of (jH)H20 formed from (3;4_3H) praline per milligram of protein per 30 minutes.
Animals receive d 150 mg/kg of dru g i.p. for four consecutive days and were s acrificed 18 hours after last injection.
Significantly different from control at P<.05. One polyvinyl sponge was implanted subcutaneously in the dorso-thoracic region of each male rat on day 1. Drug was injected directly into the sponge daily for the next 4 days and all animals were sacrificed on day 6. Enzyme activity is ex~ress e d as the mean (+ SjE.) of the amount (dpm) of ( H)H 2 o formed from (3;4-H) proline per milligram of protein per 30 minutes.
Significantly different from control at P(.05.  (Table 4). These results explain the reported increase in the amount of collagen in carrageenin induced granulomas in adrenalectomized guinea pigs (Robertson and Sanborn , 1958) and the potentiated granuloma development in adrenalectomized rats (Atkinson, et al., 1962;Ashford & Penn, 1965)   b.
Animals received 150 mg/kg of dru g i.p. for four con se cutive days and were sacrificed 18 hours after last injection. lsignificantly different fro m control at P(.05. 2significantly different from hydrocortisone at p(. 05. 3signific antly different from adrena lectomize d at P(.05. 4si gnificantly different from adrenalectomized and hydrocortisone at P(.05.

AGE OF GF A UL01 A {DAYS )
biosynthesis. Increases in proline hydroxylase activity (Figures 3, 4, 5 and 6) were observed well before literature reports of the appearance of collagen fiber deposition in growing granuloma tissue (Viljanto, 1964;Viljanto and Kulonen, 1962). produced no more than additive enzyme activity. Thus, the mechanism of increased proline hydroxylase activity in this in vivo model system in which fibroblast proliferation and collagen synthesis occurs appears not to be an activation process as is thought to be the case in growing fibroblast cells in culture (Gribble et al., 1969).
The possibility of an activation process requiring the intact tissue, such as a closely coupled system (e.g., an enzyme or biochemical substance) which functions to bring the rate limiting enzyme and proline rich polypeptide together to initiate collagen biosynthesis, cannot be ruled out. Mussini et al . (1967) observed proline hydroxylase activity on day 2 and peak activity on day 5 in carageenin induced granuloma tissue. Enzyme activity in their study was reported as specific activity (CPM/mg protein).
Fluctuations in the amount of total protein during granuloma growth were observed in the present investigation ( Figure 2). Thus, the observed increase in enzyme activity reported by Mus sini et al. (1967) may have been relative to the a mount of protein present in the 15,000 xg supernatant.
These data further support the hypothesis that proline hydroxy lase activity is a deciding factor in the formation and deposition of collagen fibers in granuloma tissue as is the case in embryonic ti s sue, foetal tissue, wound healing (Mussini et al., 1967), cc1 4 toxicity (Takeuchi et al., 1967) and the arteriosclerotic process in rabbits (Fuller and Langner, 1970). Hydroxyproline residues start accumulating after day 5 and reaches its peak at day 25 in granuloma tissue induced by sub-dermal implants of viscose cellulose sponges (Viljanto, 1964;Viljanto & Kulonen, 1962) and polyvinyl sponges (Gould, 1958). With both sponges little hydroxyproline is found up to day 5 and 6 and a few new collagen fibers are noted histologically (Viljanto, 1964;Gould, 1958). From day 7 onward accompanying the accumulation of hydroxyproline wa s noted fiber format ion with max imum accumulation of collagen fibers taking place b e t ween day 14 and day 28.
The temporal increase of proline hydroxylase activity in the cellulose sponge induced granuloma tissue could explain the early increa se in plasma hyproprotein (a collagen-like hydroxyproline containing compound) during the first week of cellulose sponge implantation observed by Kumento and Kulonen (1967). These workers reported a two-fold increase of protein-bound hydroxyproline during the accumulation of collagen in cellulose sponge induced granulomas in rats.
Since the polyvinyl spo nge posed the least problem in separating the granuloma tissue from the sponge, and the variability of enzyme activity was relatively low, this sponge type was used to screen for anti-inflammatory activity and steroidal effects on proline hydroxylase . activity. Triamcinolone and betamethasone treatment inhibi ted g r anuloma growth and decreased proline hydroxylase activity (Table 3). The data show the biochemical mechanism by which betamethasone inhibited collagen synthesis in carageenin induced inflammatory tissue (Fukuhara & Tsurufuji, 1969). Some workers (Noc enti et al., 1964) have used the argument tha t since steroids do not decre ase total hydroxypro line but do reduce collagen synthesis as det ermined by labe led proline incorporat ion, the primary steroidal effect may be to depress a b asi c c ellular metabolic process. However, steroid induced alteration of collagen resulting from decreased collagen synthesis will only be manifested when the tissue collagen has h ad time to turn over.
Although some work ers feel that the inhibitory effect of steroids on collagen may be explained by stabilization of lysosomes (Weissmann , 1965a & b) or by the increased activity of collagenase (Houck & Sharma, 1969), the results of the present investigation indicate that these drugs decrease collagen synthesis by inhibiting the rate limiting enzyme. After administration of cortisol, Houck et al. (1967) found that the los s of cutaneous collagen is at least ten-fold greater than normal loss of all tissue collagen from the whole rat.
This author . concluded that inhibition of collagen anabolism alone cannot explain the observed decrease in collagen concentration. However, in the overall synthesis of collagen propos e d by Prockop and Kivirikko (1967), the conversion of the insoluble to one of the more soluble pools resulting from a depletion of this latter pool is quite feasible. There is evidence of the clo se dependence of collagen pools on one another . The proportion of acid soluble col~agen increased in subacute inflammation produced by polyvinyl sponge implants only after the ·level of insoluble collagen remained constant (Delaunay and Bazin, 1969). Houck and Patel (1965) (Houck and Sharma, 1969) may be the mechanism for rendering the insoluble pool more soluble. However, the ability of thes e corticosteroids to inhibit proline hydroxylase probably accounts for the induced decrease in collagen content produced by these drugs.
My results are in agreement with those of Kivirikko et al. (1965) and Smith (1967) th at the corticoids decrease collagen through a depre ssant effect on the synthesis of soluble collagen. My d a ta i ndicate that the mechan ism by which cortico ids accomplish this decrease in soluble collagen is the inhibition of proline hydroxylase, the rate limiting enzyme in the collagen biosynthetic pathway.

V. SUMMARY AND CONCLUSIONS
(1) Triamcinolone, hydrocortisone and methylprednisolone significantly decreased liver praline hydroxylase activity.
A dose dependent relationship of triamcinolone on liver praline hydroxylase activity was observed.
(2) Liver praline hydroxylase activity was significantly elevated in adrenalectomized rats and was decreased to control level by hydrocortisone treatment.
(3) The temporal relationships for proline hydroxylase activity and protein content were determined during granuloma growth induced by sub-dermal implants of polyurethane, cellulose and polyvinyl sponges. The spongeinduced granuloma affords an excellent in vivo model to study the regulation of the proposed rate limiting enzyme of the collagen biosynthetic process. Studying the biochemical events temporally related to increased or decreased enzyme activity and/or the amount of enzyme in thi s model system might lead to further elucidation of the biochemical regulators of this rate limiting enzyme.
(4) The polyvinyl sponge was used to study the effects of steroids on the growth of granuloma tissue and proline hydroxylase activity. Granuloma body weight ratio was used as a parameter of anti-inflammatory activity.
Local treatment with triamcinolone and betamethasone decreased granuloma growth and inhibited praline hyd r o x ylase activity.
(5) These data support the hypothesis that the decrease in collagen content observed after glucocorticoid treatment is mediated through inhibition of proline hydroxylase activity.