Deficiencies in Pro-thyrotropin-releasing Hormone Processing and Abnormalities in Thermoregulation in Cpefat/fat Mice*

Cpefat/fat mice are obese, diabetic, and infertile. They have a mutation in carboxypeptidase E (CPE), an enzyme that converts prohormone intermediates to bioactive peptides. The Cpefat mutation leads to rapid degradation of the enzyme. To test whether pro-thyrotropin-releasing hormone (TRH) conversion to TRH involves CPE, processing was examined in theCpefat/fat mouse. Hypothalamic TRH is depressed by at least 75% compared with wild-type controls. Concentrations of pro-TRH forms are increased in homozygotes. TRH-[Gly4-Lys5-Arg6] and TRH-[Gly4-Lys5] represent approximately 45% of the total TRH-like immunoreactivity inCpefat/fat mice; they constitute ∼1% in controls. Levels of TRH-[Gly4] were depressed in homozygotes. Because the hypothalamus contains some TRH, another carboxypeptidase must be responsible for processing. Immunocytochemical studies indicate that TRH neurons contain CPE- and carboxypeptidase D-like immunoreactivity. Recombinant CPE or carboxypeptidase D can convert synthetic TRH-[Gly4-Lys5] and TRH-[Gly4-Lys5-Arg6] to TRH-[Gly4]. When Cpefat/fat mice are exposed to cold, they cannot maintain their body temperatures, and this loss is associated with hypothalamic TRH depletion and reduction in thyroid hormone. These findings demonstrate that theCpefat mutation can affect not only carboxypeptidase activity but also endoproteolysis. BecauseCpefat/fat mice cannot sustain a cold challenge, and because alterations in the hypothalamic-pituitary-thyroid axis can affect metabolism, deficits in pro-TRH processing may contribute to the obese and diabetic phenotype in these mice.

Neuropeptides and peptide hormones are first biosynthesized as precursors that must undergo a series of conversions to become biologically active (1,2). Typically, maturation of these precursors begins with limited proteolysis. Excision of the propeptide usually occurs at monobasic, dibasic, or tetrabasic residues where subtilisin-like processing enzymes cleave the precursor on the C-terminal side of these amino acids. The endoproteolysis is usually followed by the sequential removal of the basic amino acid(s) residues by a carboxypeptidase-like enzyme (3). In some situations, additional modifications can occur in the form of N-terminal acetylation or pyroglutamate formation, sulfation, and C-terminal amidation (4). These alterations usually serve to yield a peptide that is both biologically active and resistant to degradation.
Further evidence that peptide processing is physiologically relevant has come from a mouse identified at the Jackson Laboratories. This animal was reported to be obese, diabetic, and infertile because of a spontaneous mutation in the fat gene (5). Subsequent studies have shown that the fat/fat mouse has a single point mutation in the carboxypeptidase E (CPE) 1 gene, and hence, this mutation has been termed Cpe fat (6). A Ser 202 to Pro 202 transition renders the enzyme catalytically inactive and subject to rapid degradation soon after synthesis. With respect to the diabetes in the Cpe fat/fat mouse, the mechanism underlying this dysfunction has been partially attributed to a deficiency in converting pro-insulin intermediates to insulin.
To date, the mechanisms underlying the obesity in the Cpe fat/fat mouse are not well understood. The reason for these circumstances is probably because of the multifactorial nature of the regulation of food intake, absorption, and caloric utilization (7,8). In the latter case, the control of metabolism may be especially important. One hormonal system that plays an important role in the regulation of metabolism is the hypothalamic-pituitary-thyroid axis that is controlled by thyrotropin-releasing hormone (TRH).
Hypophysiotrophic pro-TRH is synthesized in the hypothalamic paraventricular nucleus, and it must undergo a number of different processing steps to yield mature TRH (9). Upon stimulation, TRH is released from median eminence nerve terminals into the hypophyseal blood where it is transported to the pituitary to stimulate the biosynthesis and secretion of thyroid-stimulating hormone (TSH; see Ref. 10). TSH, in turn, is carried in blood to the thyroid where it stimulates thyroid hormone biosynthesis and release. Besides this role, TRH can also exert some control over the release of prolactin, growth hormone, vasopressin, and insulin, as well as the classic neurotransmitters, norepinephrine and epinephrine. Furthermore, TRH is present in brain regions outside of the hypothalamus where it may serve as a neurotransmitter or neuromodulator.
In previous reports, it has been shown that the pro-TRH is cleaved endoproteolytically by two different members of the prohormone convertase (PC) family of enzymes, PC1/3 and PC2 (11). In this scheme, PC1/3 appears to be primarily responsible for most of the major cleavage events. By contrast, although PC2 can perform many of these same conversions, this enzyme is required for the formation of the pEH 24 peptide, and it is specifically involved in the processing of prepro-TRH 178 -199 to generate the pFQ 7 and pSE 14 peptides (11,12). Because PC1/3 and PC2 seem to cleave the pro-TRH on the C-terminal side of dibasic residues, these basic amino acids must be removed from the pro-TRH intermediates for the TRH to be active biologically. An enzyme that has been hypothesized to produce these conversions is CPE. Because the Cpe fat/fat mouse has a mutation in this enzyme, this animal should allow us to examine this hypothesis and determine whether this mutation exerts any effects on hypothalamic-pituitary-thyroid function.

EXPERIMENTAL PROCEDURES
Animals-The BKS.HRS-Cpe fat /J mice were received as a gift from Dr. Edward H. Leiter at Jackson Laboratories. The mouse colony was maintained at the Duke University Medical Center by heterozygous matings. Initially, mice were genotyped using the microsatellite markers D8Mit69 and D8Mit131. In these PCR, the primers were denatured at 94°C for 3 min, and the reaction was run at 94°C for 45 s, 50°C for 45 s, and 72°C for 2 min at 35 cycles. Reaction products were separated on a 3.5% 3:1 agarose gel (FMC Corp., Rockland, ME) where DNA from C57BLKS/J and B6.A-T18 a .HRS-hr mice were run as controls. Subsequently, most mice were genotyped by the procedure of Maddatu and Naggert (13). Here, DNA was denatured at 94°C for 2 min and then the reaction was run at 94°C for 30 s, 60°C for 40 s, and 72°C for 40 s at 40 cycles in a Stratagene Robo-Cycler. The PCR products were separated on a 3% MetaPhor agarose gel (FMC Corp.) where wild-type (WT) mice were identified by a 113-bp product, and Cpe fat/fat animals were identified by a 101-bp product.
The male WT, heterozygous, and homozygous mice were ϳ130 -210 days of age at the time of the study. The WT animals weighed 28.1 Ϯ 0.81 g, heterozygotes weighed 26.7 Ϯ 0.68 g, and Cpe fat/fat mice weighed 40.9 Ϯ 1.79 g. Animals were maintained on a 14:10-h light-dark cycle (lights on at 6 a.m.) in a temperature-and humidity-controlled room; food and water were provided ad libitum. Mice were sacrificed in the afternoon, and blood and the preoptic-hypothalamic area were collected (14). Hypothalami were frozen immediately in liquid nitrogen.
In the temperature regulation study, core body temperatures were taken, and mice were placed into a cold room (4 -7°C). After taking their temperatures 0, 30, 60, or 120 min later, they were sacrificed between 2 and 3 p.m. Blood and hypothalami were collected. All studies were conducted according to the principles and procedures outlined by the National Institutes of Health and by an approved protocol from the Institutional Animal Care and Use Committee at Duke University.
Peptide Extraction-The hypothalamus was dissected as described (14) and frozen in liquid nitrogen. For electrophoresis and high pressure liquid chromatography (HPLC) studies, samples were boiled for 10 min in a buffer containing 2 N acetic acid, 2 mM EDTA, 2 mM EGTA, and various enzyme inhibitors (0.1% concentrations each for phenylmethylsulfonyl fluoride, aprotinin, bacitracin, bestatin, and pepstatin) (12). Samples were sonicated, homogenized further using a Dounce homogenizer, and centrifuged at 15,000 ϫ g for 30 min. The supernatant was removed, and a small aliquot was taken from the pellet for protein analysis (15).
Chromatography-The HPLC apparatus consisted of a Hewlett-Packard 1090 system with a UV detector set to 214 nm. Peptide separations were achieved using a Vydac (Hisperia, CA) 0.45 ϫ 250-mm 218TP54 column run at 1 ml/min. In preparation for HPLC, peptide samples were filtered through Sep-Pak C 18 cartridges (Waters Associates, Milford, MA). The lyophilized samples were reconstituted in 0.05% trifluoroacetic acid and injected onto the HPLC. Initial conditions (1% methanol ϩ 0.05% trifluoroacetic acid) were held for the first 5 min, and a gradient was developed over 20 min to 3% methanol ϩ 0.05% trifluoroacetic acid. These changes in mobile phase permitted baseline separation of synthetic TRH, TRH-[Gly 4 ], TRH-[Gly 4 Lys 5 ], and TRH-[Gly 4 Lys 5 Arg 6 ] (Phoenix Pharmaceuticals, Mountain View, CA) from each other. 0.5-min fractions were collected, lyophilized, and submitted to RIA. Recovery from the column was greater than 95%.
The fractions were reconstituted with RIA buffer and screened with the TRH and TRH-[Gly 4 ] assays. Fractions that comprised and were in the vicinity of the elution positions of synthetic TRH-[Gly 4 -Lys 5 -Arg 6 ] and TRH-[Gly 4 -Lys 5 ] were incubated with recombinant carboxypeptidase D (CPD) for 4 h (see below) and submitted to the TRH-[Gly 4 ] RIA.
The RIA kit used to measure mouse TSH was developed and provided by Dr. A. F. Parlow (Harbor-UCLA Medical Center, Torrance, CA). Briefly, standards (0.195 to 50 ng of mouse TSH) and serum samples were added in a volume of 0.2 ml and were incubated with 0.1 ml of guinea pig anti-mouse TSH serum (1:8,000 dilution) for 24 h at 4°C. The next day 2% normal guinea pig serum (Antibodies Incorporated, Davis, CA) and ϳ20,000 cpm of rat [ 125 I]TSH (COVANCE, Vienna, VA) were added in a 0.1-ml aliquot and incubated for 24 h at 4°C. On the third day, anti-guinea pig sera (1:25 dilution) were added as a 0.1-ml aliquot and incubated for 24 h at 4°C. On the final day, the tubes were centrifuged at 2500 ϫ g for 30 min at 4°C, aspirated, and counted in a ␥-counter. Samples were run in duplicate. The minimal detectable dose was 0.30 ng, and the intra-and interassay variabilities were ϳ6 and 10%, respectively.
Serum triiodothyronine (T 3 ) and thyroxine (T 4 ) levels were quantitated according to the procedures and reagents provided by ICN Pharmaceuticals (Costa Mesa, CA) and Monobind (Costa Mesa, CA), respectively. The sensitivity of the T 3 and T 4 assays were 5.3 pg/ml and 0.16 ng/ml, and the intra-and interassay variabilities were ϳ5-7 and 10 -11%, respectively.
In Vitro Processing of TRH Intermediates by CPD and CPE-Baculoviral recombinants of duck CPD, rat CPE, and dansyl-Phe-Ala-Arg were generously supplied by Dr. Lloyd D. Fricker (Albert Einstein College of Medicine, Bronx, NY). For the CPD assay, 63 units (specific activity, 42 fluorescent units/g protein) of CPD were added to 100 mM Tris acetate buffer (pH 6.4) containing 2.5 mM cobalt chloride, 0.01% Triton X-100 (see Ref. 19). The CPE assay consisted of 63 units (specific activity, 49 fluorescent units/g protein) of CPE in 125 mM sodium acetate (pH 5.5), 2.5 mM cobalt chloride buffer (see Ref. 20). The solutions were incubated on ice for 15 min to activate the enzymes, and 5 g of synthetic TRH-[Gly 4 -Lys 5 ] or TRH-[Gly 4 -Lys 5 -Arg 6 ] were added. Incubations proceeded for 0.5, 4, 24, or 72 h at 37°C; at the end of this time, the samples were boiled for 5 min and frozen at Ϫ80°C. Samples were separated by HPLC (see above), and processing was quantified by UV absorbance at 214 nm.
To determine the levels of TRH-[Gly 4 -Lys 5 ] and TRH-[Gly 4 -Lys 5 -Arg 6 ] in hypothalami from WT and CPE fat/fat mice, samples were separated by HPLC, and fractions surrounding and including the elution positions for the synthetic standards were incubated with the recombinant CPD for 4 h at 37°C, boiled for 5 min, and submitted to the TRH-[Gly 4 ] RIA.
Immunocytochemistry-Experiments designed to determine whether CPE and/or CPD were expressed in TRH neurons were instituted with hypothalamic neuronal cultures. Briefly, hypothalamic neurons (3 ϫ 10 5 cells) were plated onto poly-D-lysine-coated four-chamber Lab Tek slides (Nunc, Naperville, IL). Twelve-day-old cultures were fixed with 4% paraformaldehyde in phosphate-buffered saline and submitted to immunocytochemistry as described (11,12). To identify CPE-containing neurons, cells were first incubated (1:500 dilution) with a monoclonal anti-CPE antibody (1:500; Research Diagnostics, Flanders, NY) for 24 h at 4°C. The next day, donkey anti-mouse immunoglobulin conjugated with Texas Red (1:2,000 dilution; Vector Laboratories, Burlingame, CA) was incubated with the neurons for 2 h at room temperature. For visualization of pro-TRH, slides were incubated first with anti-pAV 37 antiserum (1:1,000 dilution) at 4°C for 24 h. The next day, the slides were incubated for 2 h at room temperature with goat anti-rabbit immunoglobulin (1:500 dilution) that had been conjugated with fluorescein isothiocyanate (FITC; Vector Laboratories, Burlingame, CA). The CPD antiserum was kindly donated by Dr. Lloyd Fricker at Albert Einstein College of Medicine in Bronx, NY, and its specificity has been described (21). The CPD immunostaining was achieved by incubating (1:500 dilution) the hypothalamic neurons with the anti-CPD antiserum at 4°C for 24 h. The following day, slides were incubated for 2 h at room temperature with FITC conjugated secondary anti-rabbit sera (1:500 dilution; Vector Laboratories). In this case, the pro-TRH immunostaining was visualized with anti-pAV 37 sera that had been conjugated previously to Texas Red. Co-localization studies were performed with sequential immunostaining for CPE and pro-TRH or CPD and pro-TRH as described above. Control experiments included the incubation of cells with pre-immune sera or in the absence of the primary antibody/antiserum. Thirty-mm slides were digitized with a video camera using BioVisionframe grabbera software (Perceptice Corporation, Knoxville, TN). Images of red and green panes were combined using Adobe Systems software (Mountain View, CA) to show areas of colocalization. The objective was ϫ40.
Statistics-The data are presented as means Ϯ S.E. The data were analyzed by t tests and one-or two-way analysis of variance tests with genotype or genotype and time as the main dependent variables. A posteriori analyses were performed by Duncan and Newman-Keuls tests.

RESULTS
Hypothalamic Pro-TRH Processing-In an initial examination of pro-TRH processing, we analyzed the levels of TRH in the hypothalamus by RIA. The concentrations of immunoreactive TRH were depressed by at least 75% in the Cpe fat/fat mice (52 Ϯ 9 pg/mg protein, n ϭ 10) relative to those in the heterozygous (236 Ϯ 12 pg/mg protein, n ϭ 9) and WT controls (249 Ϯ 14 pg/mg protein, n ϭ 10). Hypothalamic protein contents did not differ among the three genotypes (data not shown). These data indicate clearly that whereas the amounts of TRH are substantially reduced in homozygote hypothalamus, some small amount of pro-TRH may be fully converted to TRH in these animals.
The reduced levels of TRH could be because of depressed biosynthesis of pro-TRH, or they could be attributed to some deficiency in processing the pro-hormone. In these studies we used antisera that were raised against the rat pro-TRH and TRH sequences, because no mouse-specific reagents were available. It should be emphasized that the amino acid sequence homology between the two species is ϳ88% with 100% conservation around the processing sites for the five cryptic TRH sequences (see Refs. 22 and 23, and see Fig. 1). In these experiments, we used anti-TRH, anti-TRH-[Gly 4 ], and anti-pCC 10 serum. The latter antiserum was generated against the synthetic decapeptide, Cys-Lys-Arg-Gln-His-Pro-Gly-Lys-Arg-Cys (where the underlined amino acids represents the TRH sequence), and it binds to the pro-TRH (26 kDa) and the 15-, 9.5-, and 6-kDa intermediate peptides (11). Separation of pro-TRH and its high molecular mass intermediates by SDS-PAGE revealed that the precursor was processed into a number of different intermediates (Fig. 2). Indeed, six peptides unidentified previously were detected in the hypothalami from WT and Cpe fat/fat mice using the anti-pCC 10 serum. Interestingly, five of these peptides (15, 11, 7.5, 5, and 4 kDa) were increased substantially in the homozygotes over those of the controls (Table I). In both genotypes, the majority of the immunoreactivity was found in the 1.6-kDa species. Collectively, these data indicate that processing of the pro-TRH is aberrant in hypothalami from Cpe fat/fat animals. Moreover, they also demon-strate that in homozygotes the overall quantities of the pro-TRH intermediates are somewhat higher in this brain region than those from the WT littermates.  (22,23). The rat sequence is 255 amino acids in length, whereas the mouse sequence is 256. When the sequences are aligned, the mouse sequence is found to have an insertion at amino acid 194. There is ϳ88% homology between the sequences.

FIG. 2. Electrophoretic separation of hypothalamic high molecular mass pro-TRH-derived peptides from male WT and
Cpe fat/fat mice. Proteins were separated by SDS-PAGE, eluted from the gel, and submitted to RIA using the anti-pCC 10 serum. Top, separation from WT hypothalamus. Bottom, separation from mutant hypothalamus. Recovery from the gel was ϳ90%, and each data point was replicated twice. The sizes of the pro-TRH species were estimated by molecular mass standards.
Because conversion of pro-TRH to its constituent high molecular mass intermediates is perturbed in hypothalamus of Cpe fat/fat mice, we next examined whether processing of the low molecular mass peptides is also abnormal. When the HPLC fractions were screened with either the TRH or the TRH-[Gly 4 ] antisera, each of the antisera only detected a single immunoreactive peak, and these materials were found to elute in the same positions as the respective synthetic peptide standards (Fig. 3). TRH concentrations were reduced ϳ4-fold in hypothalami from the Cpe fat/fat animals compared with their WT littermates (Table II). Although some TRH-[Gly 4 ] was present in WT samples, the concentrations of this peptide were reduced by ϳ10-fold in the mutants.
To quantitate the concentrations of the TRH-[Gly 4 -Lys 5 ] and TRH-[Gly 4 -Lys 5 -Arg 6 ], fractions corresponding to the elution positions of the synthetic standards were incubated with recombinant CPD to convert the endogenous peptides to TRH-[Gly 4 ], and these peptides were quantitated by RIA (Fig. 3). When the data are expressed as a mol percent of the total TRH-like immunoreactivity (IR), the amounts of TRH-[Gly 4 -Lys 5 -Arg 6 ] are increased 60-fold whereas those for TRH-[Gly 4 -Lys 5 ] are enhanced 3-fold in the mutant relative to WT hypothalamus (Table II). This same condition can be seen readily when the results are expressed as the molar ratio of each of the low molecular mass TRH intermediates relative to that for TRH. These data clearly indicate that removal of the basic amino acids from the C terminus of the low molecular mass pro-TRH intermediates is deficient in the Cpe fat/fat mouse.
Immunocytochemistry-The low molecular mass TRH intermediates contain basic residues at their C termini. Presumably, these amino acids are removed by CPE. To determine whether this carboxypeptidase is expressed in TRH neurons, co-localization studies were employed in our well established primary hypothalamic cultures (11,12). CPE-like IR can be seen (red) in both the cell body regions and in the processes of these neurons (Fig. 4A). As expected, pro-TRH-like IR (green) is confined almost exclusively to the perikarya (Fig. 4B). Overlay of these immunoreactivities (orange) reveals that CPE and pro-TRH are present within the same neurons (Fig. 4C).
In the Cpe fat/fat mouse, the enzyme is unstable, and it is degraded in the endoplasmic reticulum (6,24). Despite this fact, although mutation of the CPE gene depresses processing of TRH intermediate peptides dramatically, some TRH is found in the hypothalami of these animals. These data suggest that a carboxypeptidase besides CPE may process the low molecular mass TRH intermediates in these neurons. One such carboxypeptidase that has been found to reside in many different tissues, including brain, is CPD (21,25). To determine whether this carboxypeptidase normally resides in TRH neurons, we used the primary hypothalamic neuronal cultures. CPD-like IR (green) can be visualized to be distributed almost exclusively in the perikarya of the neurons (Fig. 4D). Similarly, TRH-like IR (red) is also present in the cell body (Fig. 4E). Co-localization of the immunostaining demonstrates that CPD-like and TRH-like immunoreactivities (orange-yellow) reside within the same neurons (Fig. 4F). The pro-TRH, CPE, and CPD immunostainings are specific, because no immunoreactivity is detected for any of these antigens when pre-immune sera (for pro-TRH) is used or when the secondary antisera/antibody (for CPE and CPD) is omitted from the reaction (Fig. 4, G-I). These data support the idea that a carboxypeptidase besides CPE is present in TRH neurons.
In Vitro Processing of TRH Intermediates-To evaluate whether CPE or CPD can process the low molecular mass TRH intermediates, synthetic TRH-[Gly 4 -Lys 5 ] or TRH-[Gly 4 -Lys 5 -Arg 6 ] were incubated with recombinant CPE or CPD using identical units of enzyme. During the first 30 min of incubation with CPE, only ϳ10% of the TRH-[Gly 4 -Lys 5 -Arg 6 ] is converted to TRH-[Gly 4 -Lys 5 ] (Fig. 5A)  a Percent refers to the mole percent of total pro-TRH products that is contributed for a given pro-TRH product.
b Mol % refers to the femtomol ratio of a given peptide product for Cpe fat/fat mice compared to that same product for WT animals times 100.
c WT/fat MR refers to the femtomol ratio of peptide product for WT and Cpe fat/fat mice.  4 ] antisera. Bottom, a chromatographic profile of materials from homozygous mice. Recoveries were ϳ95%, and each data point was replicated three times. Although each original chromatogram was run with at least 10 mice, the values in the chromatogram are presented in terms of those for an individual animal. (Fig. 5B). These data suggest that removal of the Lys 5 residue by the CPE may be more protracted than excision of the Arg 6 from the intermediate. By comparison, CPD processing of TRH-[Gly 4 -Lys 5 -Arg 6 ] to TRH-[Gly 4 -Lys 5 ] and TRH-[Gly 4 -Lys 5 ] to TRH-[Gly 4 ] occurs very rapidly and is essentially complete within the first 30 min of incubation (Fig. 5, C and D). Collectively, these data indicate that both CPE and CPD can process TRH intermediates in vitro.
Status of the Hypothalamic-Pituitary-Thyroid Axis in the Cpe fat/fat Mice-Because processing of the pro-TRH is deficient, and TRH levels are reduced substantially in hypothalami from male Cpe fat/fat mice, concentrations of TSH, T 4 , and T 3 were measured in serum. Baseline levels of serum TSH (Fig. 6A) and T 4 (Fig. 6B) were not distinguished by genotype. By contrast, circulating T 3 values were significantly higher in homozygote blood (Fig. 6C). These data demonstrate that, for the most part, mutation of the Cpe gene does not substantially affect pituitary or thyroid function under basal conditions.
To study the dynamics of hypothalamic-pituitary-thyroid interactions, mice were placed into a cold room (4 -7°C) for 0, 30, 60, or 120 min. At the outset of the experiment (0 min), the core body temperatures in the mutants were somewhat lower (ϳ0.6°C) than those in the WT animals. Controls were able to maintain their core temperatures, with little fluctuation (increase of 0.2°C) over the 2-h period (Fig. 7A). By contrast, within the first 30 min of cold exposure, core body temperatures in mutants declined significantly, and by the end of 2 h, their body temperatures were reduced by at least 3°C (two mice showed a reduction by 6°C). Because the hypothalamicpituitary-thyroid axis is involved intimately in thermoregulation (26), these data suggest that this axis in the Cpe fat/fat mice is deficient.
Assessments of hormonal function were also performed during the 2-h exposure to cold. Hypothalamic TRH values in homozygous mice were already depressed at the zero time point by ϳ79% compared with their WT littermates (Fig. 7B). Con-

FIG. 4. Co-localization of CPE-and CPD-like IR in hypothalamic TRH
neurons. Primary hypothalamic cultures were cultured for up to 14 days, fixed with 4% paraformaldehyde, and immunostained for CPE, CPD, and TRH. A, CPE immunoreactive perikaya (arrows) and processes (arrows) revealed with an antiserum conjugated to Texas Red. B, pro-TRH immunoreactive cell bodies (arrows) as visualized with FITC-conjugated antiserum. C, co-localization of CPE-and pro-TRH-like IR (arrowheads) within the same neurons. D, CPD immunoreactive neurons (arrows) detected with FITC-conjugated antisera. E, pro-TRH immunoreactive neurons (arrows) revealed with Texas Red antisera. F, co-localization of CPD-and pro-TRH-like immunoreactivity (arrowheads). G, pro-TRH immunoreactivity with pre-immune serum. H, CPE immunostaining in the absence of the primary antiserum. I, CPD immunoreactivity in the absence of primary antiserum. Objective, ϫ40. tinual exposure to cold over the next 2 h depressed hypothalamic TRH levels significantly in the mutants by ϳ38%, whereas those in the WT animals were only attenuated by ϳ13%. An examination of TRH-[Gly 4 ] levels revealed that these values did not change over the 2-h period for either genotype (data not shown). Because processing of the Arg 6 -and Lys 5 -extended intermediates is blocked in the mutant, these findings suggest that the TRH-Gly 4 is not readily available for secretion. Together, these data indicate that cold exposure reduces hypothalamic TRH contents in both genotypes and that the stores of this peptide are reduced more dramatically in the homozygotes than in the WT mice.
Exposure to cold increased serum TSH levels in both genotypes; however, the response in the Cpe fat/fat animals was initially more robust than that in the WT controls (Fig. 7C). This significantly increased response could not be maintained in the mutants, and serum TSH values declined significantly to a value that was 83% from baseline. By contrast, in WT animals there was a tendency for serum TSH contents to be increased by cold, but by the end of the 2-h period these values were reduced by less than 20%. These data suggest that the Cpe fat/fast mice may be able to defend their temperatures during brief exposure to cold, but more prolonged conditions compromise the ability of the hypothalamic-pituitary axis to mount a sustained response.
An examination of serum T 4 concentrations revealed that these values effectively mirror those for TSH (Fig. 7D). Hence, serum T 4 levels in homozygotes were increased during the first 30 min in the cold environment; however, they declined over time. By comparison, the WT mice showed a more protracted response to cold, and the levels of this hormone were increased and maintained over the 2-h period of exposure. Hence, at the end of this time, the concentration of T 4 was significantly higher in sera from the WT animals compared with that for the Cpe fat/fat mice.
As previously noted, basal serum T 3 contents are already elevated in the Cpe fat/fat animals relative to their WT controls. Upon cold exposure, the mutants showed no further increase in serum T 3 levels; instead, values declined significantly by ϳ30% over the 2-h period (Fig. 7E). By comparison, WT animals responded with significant increase in circulating T 3 levels to the cold during the first 30 min of cold exposure, with a decline to baseline values by 120 min. Collectively, these data indicate that although both genotypes of mice show endocrine responses to the cold, those from the mutants are transient and are not sufficient to maintain their core body temperatures. DISCUSSION CPE is an exopeptidase that is responsible for removing basic amino acids from the C-terminal of proteins and peptides (3). A point mutation in the CPE gene (e.g. Ser 202 to Pro 202 transition) is sufficient to reduce the efficiency of endoproteolytic cleavage of the pro-TRH and its high molecular mass products in Cpe fat/fat hypothalamus. These data indicate that mutation of this gene can exert effects upstream of its normal processing activity. In rats, pro-TRH is processed by PC 1/3 and 2 (11,12). Both endoproteases are synthesized as pro-enzymes and must be processed to attain full activity (1, 2). It has been proposed that CPE may participate in the full activation of these convertases through removal of C-terminal basic amino acids or by inactivation of endogenous inhibitors (27)(28)(29). Recently, it has been reported that protein levels and enzymatic activities of both enzymes are altered in the brains of the Cpe fat/fat mice (30). As a consequence, levels of pro-dynorphin and its high molecular mass intermediate (e.g. dynorphin A-17) are increased (30,31). Additionally, concentrations of pro-insulin in pancreas (6) and pro-opiomelanocortin in pituitary are enhanced (29,31,32). Despite these findings, impaired endoproteolysis of pro-peptides is not necessarily a common feature of the CPE mutation, because the levels of pro-cholecystokinin in brain are unchanged (33,34), whereas those in intestine are reported to be either enhanced (34) or unaltered (33) from the WT controls.
Besides alterations in TRH levels, concentrations of some high and low molecular mass pro-TRH intermediates were also changed in the Cpe fat/fat hypothalamus. Although the function of TRH has been studied for many years, only recently have physiological roles been ascribed to some of the other pro-TRH products (9). For instance, although 100-fold less potent than TRH, TRH-Gly 4 stimulates gastric acid secretion in a dose-dependent manner (35). Additional peptides that have biological activity include the prepro-TRH160 -169 and the prepro-TRH178 -199 (36,37). The prepro-TRH160 -169 augments TRH-stimulated TSH secretion and potentiates TRH-induced gastric acid secretion when microinjected into the dorsal motor nucleus of the vagus nerve. The prepro-TRH178 -199 can serve as a corticotropin-releasing hormone inhibiting factor (37,38), and it can stimulate prolactin release from the pituitary (12).
Inasmuch as these pro-TRH-derived peptides possess biological activity, and because processing of the pro-TRH intermediates are affected by the CPE mutation in mice, it may be expected that some of these functions will be abnormal in the Cpe fat/fat mouse.
In the present study, hypothalami from Cpe fat/fat mice were found to contain high levels of the low molecular mass intermediates, TRH-[Gly 4 -Lys 5 -Arg 6 ], and TRH-[Gly 4 -Lys 5 ]. These data indicate that carboxypeptidase activity is deficient in vivo and that CPE is primarily responsible for this activity in TRH neurons. A role for CPE in TRH processing is strengthened further by the in vitro conversion of TRH-[Gly 4 -Lys 5 -Arg 6 ] and TRH-[Gly 4 -Lys 5 ] to TRH-[Gly 4 ] and by the co-localization of CPE-like IR in TRH neurons. It should be noted that besides TRH intermediates, an increase in additional C-terminal basic amino acid-extended peptides has been observed for insulin (6), cholecystokinin (33,34), gastrin (39,40), and neurotensin in the Cpe fat mutant (41).
In murine hypothalamus, the Cpe fat mutation depresses the levels of fully processed TRH. Additional peptides in brain that are also reduced include dynorphin (30,31), cholecystokinin (33), neurotensin (41), and substance P (42). Besides removal of basic amino acids, amidation may also be affected in the Cpe fat/fat mouse. For instance, the molar ratio of TRH-[Gly 4 ] to TRH and the percent of this glycine-extended peptide relative to the other TRH-like low molecular mass peptides were depressed in mutant hypothalami. A similar relationship has also been reported for glycine-extended cholecystokinin (34) but not for gastrin (40) in gut. Together, these data suggest that amidative activity may be influenced in the Cpe fat/fat mouse.
Despite perturbations in pro-TRH processing, Cpe fat/fat hypothalamus contains low quantities of TRH. This result suggests either that some residual CPE is present or that an additional carboxypeptidase resides in TRH neurons. One possible candidate is CPD. This enzyme is expressed in many brain regions including the hypothalamic paraventricular nucleus where the hypophyseal TRH neurons are located (25). Moreover, CPD is co-localized with CPE in many different brain regions. In the present report, we show that hypothalamic TRH neurons contain CPE-and CPD-like IR. Moreover, recombinant CPD can process synthetic TRH-[Gly 4 -Lys 5 -Arg 6 ] or TRH-[Gly 4 -Lys 5 ] in vitro. These findings clearly show that CPD can mimic the actions of CPE. Despite this fact, hypothalamic levels of TRH-[Gly 4 -Lys 5 -Arg 6 ] and TRH-[Gly 4 -Lys 5 ] are augmented in Cpe fat/fat mice. One reason for the discrepancy between the in vitro and in vivo results may be the differential locations of the enzymes. Our immunocytochemistry results show that CPE-like IR is located in the perikarya and processes, whereas the CPD-like IR is confined to the cell body region of TRH neurons. These findings complement other studies where CPE has been localized primarily to mature secretory granules (43), whereas CPD resides in the trans-Golgi network and immature secretory granules (44). Because pro-TRH processing begins in the Golgi and continues in the secretory granule (16,45), CPD could begin conversion of the pro-TRH intermediates in vivo; however, these effects would be transient as the intermediates passed to the mature granule and a build-up of intermediates could occur.
Cold exposure stimulates the hypothalamic-pituitary-thyroid axis strongly (26). In rats, cold temperatures increase TRH mRNA levels in the hypothalamic paraventricular nucleus (46), augment TRH release from the median eminence (47), stimulate TSH secretion from pituitary, and potentiate T 4 and T 3 contents in blood (48). In our studies, basal levels of serum TSH and T 4 are similar among WT, heterozygous, and Cpe fat/fat mice. A similar relationship has been reported for serum T 4 concentrations in homozygotes and lean controls (7). Although Cpe fat/fat animals respond to cold, they cannot maintain their core body temperatures or sustain their endocrine responses. The already depressed hypothalamic TRH levels in the Cpe fat/fat mice are reduced further by an additional 38% with cold exposure. Although TRH neurons reside in several different hypothalamic areas, only the paraventricular nucleus controls anterior pituitary function (50). The large reduction of TRH stores in response to cold suggests that much of the fully processed TRH is located in a readily releasable pool of peptide and that substantial quantities of TRH are released from the hypophyseal neurons. Together, these findings demonstrate that the hypothalamic-pituitary-thyroid axis is impaired in its response to cold and that the TRH processing deficit in the Cpe fat/fat mouse may contribute to this deficiency.
The fat/fat mouse was identified initially at Jackson Laboratories as being obese, diabetic, and infertile (5). Although the regulation of each of these physiological processes is complex, the deficiency in pro-TRH conversion to TRH may contribute to its obesity and diabetes. For instance, disruption of the TRH gene in mice produces alterations in insulin secretion and hyperglycemia (51). Thyroid hormone influences oxygen consumption and metabolism in mammals, and changes in hypothalamic-pituitary-thyroid function can affect weight gain and appetite (52,53). In addition, the TSH response to TRH stim-ulation is increased in obesity (49). In our studies the TSH response to cold exposure was more robust in the mutants than in the WT animals. This enhanced responsivity in the presence of reduced hypothalamic tissue stores of TRH suggests that TRH levels in hypophyseal blood may be low in the Cpe fat/fat mice and that the TRH receptor may be up-regulated on the pituitary. Similar to cold exposure, these circumstances could render this mutant unable to respond normally to metabolic and other challenges. Inasmuch as TRH is found in pancreas, gastrointestinal tract, pineal gland, neurohypophysis, and many regions of the central nervous system (10), a deficiency in pro-TRH processing could affect the physiology in a number of additional organs and systems.