The Effects of 20-kDa Human Placental GH in Male and Female GH-deficient Mice: An Improved Human GH?

Abstract

A rare 20K isoform of GH-V (here abbreviated as GHv) was discovered in 1998. To date, only 1 research article has characterized this isoform in vivo, observing that GHv treatment in male high-fat fed rats had several GH-like activities, but unlike GH lacked diabetogenic and lactogenic activities and failed to increase IGF-1 or body length. Therefore, the current study was conducted to further characterize the in vivo activities of GHv in a separate species and in a GH-deficient model (GH-/- mice) and with both sexes represented. GHv-treated GH-/- mice had significant increases to serum IGF-1, femur length, body length, body weight, and lean body mass and reduced body fat mass similar to mice receiving GH treatment. GH treatment increased circulating insulin levels and impaired insulin sensitivity; in contrast, both measures were unchanged in GHv-treated mice. Since GHv lacks prolactin receptor (PRLR) binding activity, we tested the ability of GH and GHv to stimulate the proliferation of human cancer cell lines and found that GHv has a decreased proliferative response in cancers with high PRLR. Our findings demonstrate that GHv can stimulate insulin-like growth factor-1 and subsequent longitudinal body growth in GH-deficient mice similar to GH, but unlike GH, GHv promoted growth without inhibiting insulin action and without promoting the growth of PRLR-positive cancers in vitro. Thus, GHv may represent improvements to current GH therapies especially for individuals at risk for metabolic syndrome or PRLR-positive cancers.

Growth hormone (GH) is a 22-kDa protein secreted by the anterior pituitary. While it is named for its role in controlling somatic growth, GH exerts a multitude of other effects, influencing most tissues and organs either directly or indirectly via secreted factors. In humans and other mammals, insulin-like growth factor-1 (IGF-1) is a major mediator of GH action. Growth hormone also affects nutrient partitioning, impacting the oxidation, synthesis, and distribution of proteins, carbohydrates, and fats. In fact, one of the first described actions of GH is its diabetogenic or anti-insulin activity. This diabetogenic activity was demonstrated nearly a century ago by Bernardo Houssay, who showed that surgical removal of the pituitary in dogs and toads increased insulin sensitivity, and exogenous treatment of pituitary extracts causes insulin resistance and diabetes (1, 2). Houssay later received the Nobel Prize in 1947 for this work. Subsequently, experiments by Rabinowitz in the early 1960s showed that GH is a diabetogenic agent in humans, as it robustly blocked insulin-stimulated glucose uptake when GH and insulin were coadministered (3, 4). Since then, numerous studies in multiple species have confirmed that GH has potent diabetogenic activities (5–9). Therefore, the development of a GH that promotes growth with diminished diabetogenic activity represents a potential improvement to current GH therapies.

In humans, the GH gene family consists of 5 genes: (1) GH (also called GH-N or GH1), (2) placental GH (also referred to as GH-V or GH2), (3) chorionic somatomammotropin hormone 1, (4) chorionic somatomammotropin hormone 2, and (5) a pseudogene called chorionic somatomammotropin-like hormone. The chorionic somatomammotropin hormones are also referred to as placental lactogens. All of these genes share a similar structure and are believed to have evolved from a common ancestor (10). Unlike GH, which is expressed primarily but not exclusively in the pituitary, GH-V is produced in the syncytiotrophoblast cells of the placenta, fundamentally replacing pituitary GH during pregnancy. GH-V differs from pituitary GH in 13 of the 191 amino acid residues (10, 11). The most abundant form of GH-V, the 22K isoform, has been shown to promote growth and mediate maternal insulin resistance. In 1998, a 20kDa GH-V isoform was discovered by Cesar Boguszewski in human placenta (12). The 20K version of GH-V results from a 45-bp deletion produced via the use of an alternative acceptor site within exon 3. Unfortunately, very little is known about the activities of 20K GH-V, as only 2 papers have evaluated this isoform in vitro (11) or in vivo (13). In vitro studies by Solomon indicate that 20K GH-V has a very low affinity for the prolactin receptor (PRLR), suggesting that this isoform lacks the lactogenic activity found in “normal” human GH (11). In the only in vivo study, Vickers et al report that 20K GH-V treatment in high-fat fed rats produces several GH-like activities, such as reduced body fat and increased body weight, but fails to stimulate increases in serum IGF-1, body length, or insulin (13). Since GH-stimulated IGF-1 is essential for longitudinal bone growth, the fact that 20K GH-V also failed to stimulate increases in IGF-1 or body length may have dampened the enthusiasm for potential clinical utility. As such, no studies have since been published on the 20K GH-V. However, Bielohuby and colleagues report that normal (non-GH deficient) wild-type laboratory rodents have a limited IGF-1 response to exogenous GH treatment despite dramatic changes to glucose homeostasis, body composition, and organ mass (14, 15). Bielohuby speculates that IGF-1 release may be near maximum in these animals, making further increases by exogenous GH treatments difficult to achieve (14). Thus, the use of GH-deficient models may be more appropriate for evaluating GH action via exogenous treatment. Accordingly, it remains uncertain whether 20K GH-V can stimulate IGF-1 and if it has utility as a therapeutic for GH replacement. Considering the limited knowledge of 20K GH-V action in vivo and the potential benefits of a GH with diminished diabetogenic and lactogenic activities, the purpose of this study was to compare the various activities of 20K GH-V versus 22K GH-N (which we refer to as GH) using male and female GH-deficient (GH-/-) mice.

Materials and Methods

GH-/- mice and GH treatments

GH-/- mice used in this study have been described previously (16). GH-/- mice were housed 3 to 4 per cage and given ad libitum access to water and rodent chow (ProLab RMH 3000; 14% of energy from fat, 60% from carbohydrates, and 26% from protein; PMI Nutrition International, Brentwood, Missouri). The cages were maintained in a temperature- (22°C) and humidity-controlled room and exposed to a 14-hour light, 10-hour dark cycle. For all mouse studies, 4-week-old male and female GH-/- mice were injected subcutaneously in the nape at the back of the neck with 5 ug/g of human GH-N 22K (which we subsequently will refer to as GH) or human GH-V 20K (which we refer to as GHv) or saline, once daily for 6 days. The 2 forms of GH (hGH-N 22K, Cat no. GHP-24 and hGHv 20K, Cat no. GHP-12) were purchased from Protein Laboratories Rehovot (Rehovot, Israel). The GH from this company is produced in Escherichia coli and purified, as previously described (11). Both forms of GH were reconstituted in 0.4% NaHCO3 and mixed by gentle inversion to a concentration of 500 ug/ml and were administered by subcutaneous injection. This dose of GH was based on previous dosing studies in C57BL/6J mice required for observable changes to body composition and IGF-1 (17), and it was also the same dosing used by Vickers et al (13). GH-/- mice were generated by crossing heterozygous GH+/- males to heterozygous GH+/- females to produce 113 litters with 662 offspring, of which 76 were homozygous GH-/-. These 76 GH-/- mice were randomly distributed into 6 experimental groups with 12 to 14 mice per group per sex used for GH treatment studies, except for microcomputed tomography, where a subset (n = 6–7) were analyzed. All procedures were approved by the Ohio University Institutional Animal Care and Use Committee and fully complied with all federal, state, and local policies.

Body weight and body composition measurements

Body weight was measured using a Mettler Toledo PL 202-S balance. Body composition was measured using the Minispec mq Benchtop Nuclear Magnetic Resonance (NMR) analyzer (Bruker Instruments, Minispec ND2506), as previously described (17–19).

Fasting blood glucose and insulin tolerance tests

These procedures have been described previously (20, 21). Briefly, fasting blood glucose was determined following a 6-hour fast using OneTouch Ultra test strips and glucometers (Lifescan, Inc, Milpitas, California). Blood samples were obtained by cutting approximately 1 mm from the tip of the tail and collecting the first drop of blood. Insulin tolerance tests (ITTs) were performed following a 6-hour fast. Recombinant human insulin (Humulin-R; Eli Lilly & Co, Indianapolis, Indiana) was prepared by diluting Humulin-R (100 U/ml) to 0.075 U/mL in sterile 0.9% NaCl. Each mouse received an i.p. injection of the 0.075 U/ml insulin solution at a dose of 0.75 U/kg body weight. Blood glucose was measured before the insulin injection and at 15, 30, 45, 60, and 90 minutes after injection.

Body length, bone length, blood collection, and tissue collection

At the end of the GH treatment period, mice were sacrificed following a 12-hour overnight fast. Mice were placed in a CO₂ chamber until unconscious, and then blood was collected from the orbital sinus. Serum was separated from blood and stored at -80°C until assays could be performed. Body length was determined prior to tissue collection, measuring the length from nose to anus. In order to determine femur length, the right hind limb was dissected at the hip followed by removal of soft tissues from the bone. Femora were then measured using micrometer calipers (0.01 mm divisions Item # 601A, Walter Stern, Inc. Port Washington, New York). A subset of the left hind limbs (n = 6–7) were also collected and shipped to Dr Shoshana Yakar for determination of bone mineral density and strength by microcomputed tomography (described in the next section). Adipose tissues (inguinal subcutaneous, perigonadal, retroperitoneal, mesenteric, and interscapular brown fat), liver, kidney, heart, lung, spleen, brain, skeletal muscle (gastrocnemius, soleus, and quadriceps) were collected, weighed, and flash frozen in liquid nitrogen and then stored at -80°C or processed for current or future analysis.

Micro-computed tomography

Micro-computed tomography (micro-CT) analyses were performed by Dr Shoshana Yakar at the New York University, College of Dentistry according to the guidelines published by The American Society for Bone and Mineral Research (22). The left femora were scanned using a high-resolution SkyScan micro-CT system (SkyScan 1172, Kontich, Belgium). Images were acquired using a 10 MP digital detector, 10W energy (100kV and 100 mA), and a 0.5 mm aluminum filter with a 9.7 μm image voxel size. A fixed global threshold method was used based on the manufacturer’s recommendations and preliminary studies, which showed that mineral variation among groups was not high enough to warrant adaptive thresholds. The cortical region of interest was selected as the 2.0 mm mid-diaphyseal region directly below the third trochanter, which includes the mid-diaphysis and more proximal cortical regions. The trabecular measurements were taken at the femur distal metaphysis 0.25 mm below the growth plate.

Serum measurements

IGF-1 levels were measured using an IGF-1 mouse/rat ELISA (22-IG1MS-E01; ALPCO). Insulin and C-peptide serum levels were measured using a Milliplex Mouse Metabolic Panel (MMHMAG-44K; Millipore). Prolactin, thyroid-stimulating hormone (TSH), follicle-stimulating hormone (FSH), and adrenocorticotropic hormone (ACTH) serum levels were measured using a Milliplex Mouse Pituitary Panel (MPTMAG-49K). All Milliplex panels were analyzed using a Milliplex 200 Analyzer (Millipore, Burlington, Massachusetts) according to the manufacturer’s instructions and as previously described (18).

Reverse transcription and quantitative polymerase chain reaction

Six distinct human cancer cell lines (described below) were used to compare their ability to respond to GH or prolactin by measuring their respective receptor expression levels. Total RNA isolation from human cancer cells was done using a Trizol-based Total RNA purification kit (IBI Scientific, St. Louis, Missouri), and 1 ug of total RNA was reverse transcribed using the Maxima First Strand cDNA synthesis kit (Thermo Fisher Scientific, Waltham, Massachusetts) following the manufacturer’s protocols. A total of 3 uL of 1/15 diluted cDNA was used in a 20 uL quantitative polymerase chain reaction (PCR) using Maxima SYBR-Green qPCR master mix (Thermo Fisher Scientific) and a T100 thermal cycler (Biorad, Hercules, California). RNA and DNA concentrations were estimated using NanoDrop 2000 (Thermo Fisher Scientific) spectrophotometer. Primers were obtained from Sigma-Aldrich for the following human genes, and primer efficiencies were confirmed. Amplification of specific transcripts was confirmed by the melting-curve profiles (cooling the sample to 68°C and heating slowly to 95°C, with the measurement of fluorescence) at the end of each PCR. Primer sequences used included: GHR forward 5’-CTCCTCAAGGAAGGAAAATTAG-3’; GHR reverse 5’-GTGGAATTCGGGTTTATAGC-3’; PRLR forward 5’-CAGAGTTTAAGATTCTCAGCC-3’; PRLR reverse 5’-TGGTGAAGTCACTAGGTATC-3’; ACTB (reference gene) forward 5’-GACGACATGGAGAAAATCTG-3’; and ACTB reverse 5’-ATGATCTGGGTCATCTTCTC-3’.

Cell proliferation/viability assays

Six different human cancer cell lines were used to compare the ability of GH and GHv to stimulate proliferation. The cells used were: T84 (colon adenocarcinoma; ATCC Cat# CCL-248, RRID:CVCL_0555) (23), T47D (breast/invasive ductal carcinoma; ATCC Cat# HTB-133, RRID:CVCL_0553) (24), MDA-MB-231 (breast adenocarcinoma; ATCC Cat# HTB-26MET, RRID:CVCL_VR67) (25), A498 (renal cell carcinoma; ATCC Cat# CRL-7908, RRID:CVCL_1056) (26), H1299 (non-small cell lung carcinoma; ATCC Cat# CRL-5803, RRID:CVCL_0060) (27), and SK-MEL-28 (cutaneous melanoma; ATCC Cat# HTB-72, RRID:CVCL_0526) (28). All cells were purchased and validated from American Type Culture Collection (Manassas, VA) and maintained in a respective cell culture media containing 1X-penicillin-streptomycin and 10% fetal bovine serum (FBS). For cell proliferation/viability assays, cells were plated in 96-well plates at 10 000 cells per well and, following cell attachment (16–18 hours), either GH or GHv were added at equimolar proportions at 0.6, 2.5, and 10 nM concentrations in 2% FBS-containing media and incubated for 48-hours in a cell-culture incubator at 37°C. Finally, cell viability was assessed using PrestoBlue assay (Invitrogen, Carlsbad, CA), following the manufacturer’s protocol.

Statistics

Values are given as means ± SEM. Assumptions of normal distributions were verified by histograms and quantile-quantile plots. Statistical analysis was performed using GraphPad Prism version 7.01 version (GraphPad Software, Inc., CA). For all comparisons except those noted below, 2-way ANOVA tests were used followed by Tukey’s post hoc test. For comparison of longitudinal data (body weight, fat mass, lean mass, and fluid mass over time), repeated measures ANOVA was used. For micro-CT bone measures, since no drug effect, sex effect, or interaction effect was observed by the 2-way ANOVA, male and female data were combined to increase N, and thus 1-way ANOVA was performed with Tukey’s post hoc analyses. The significance level (alpha) was set to 0.05 in all cases.

Results

GHv alters body weight, lean mass, fat mass, and fluid mass

Growth hormone has potent effects on body composition, being catabolic in adipose tissue and anabolic in skeletal muscle, bone, and most other tissues (29). Therefore, to test the ability of GHv to stimulate changes in body composition, we used NMR to measure body composition just prior to the first injection and once daily during injections. Both GH and GHv treatments significantly increased body weight (Fig. 1A and 1B) in male and female GH-/- mice compared to saline-treated controls. Similarly, lean mass (Fig. 1C and 1D) and fluid mass (Fig. 1G and 1H) were significantly increased in both sexes, with both GH and GHv treatments compared to controls. The opposite was observed for body fat (Fig. 1E and 1F), as both GH and GHv treatments significantly decreased fat mass in both sexes compared to controls. For all body composition measurements, GH and GHv treatment did not differ from one another.

Figure 1.

Both GH and GHv increase body weight, lean mass, and fluid mass, and decrease fat mass. Daily body weights over 6 days of treatment are shown for male (A) and female (B) GH-/- mice treated with saline (black), GH (red), or GHv (green). Lean body mass (C and D), fat mass (E and F), and fluid mass (G and H) data are shown for male and female mice, respectively. N = 12–14 per group per sex for all measures. # indicates that both GH and GHv are significantly different (P < 0.05) from saline controls and that GH and GHv do not differ from one another, as determined by repeated measures ANOVA.

GHv stimulates longitudinal growth

Longitudinal bone growth is regulated by GH in a dose-dependent manner. To test the ability of GHv to stimulate longitudinal bone growth, we measured serum IGF-1, femur length, and body length in mice after six days of treatment. Serum IGF-1 was significantly increased in both sexes (Fig. 2A), with both GH and GHv treatments compared to saline-treated controls. IGF-1 levels did not differ between GH or GHv treatments in males but was higher with GHv treatment in females. Body length (Fig. 2B) and femur length (Fig. 2C) were significantly increased in both sexes, with both GH and GHv treatments compared to saline-treated controls, although GH and GHv treatments did not differ from one another. In order to evaluate bone morphology, a subset of femurs were sent to the NYU College of Dentistry (New York) for micro-CT analysis, where no observable changes were detected in all measures, including cortical bone mineral density, cortical strength (as measured by moment of inertia polar), or trabecular bone mineral density even when male and female data were grouped together (Table 1).

Figure 2.

Both GH and GHv increase serum IGF-1, femur length, and body length. A: Serum IGF-1 levels are shown for male and female GH-/- mice. B: Nasal-anal body length and femur length (C) are shown. Samples were collected from 5-week-old mice following 6 days of treatment with saline (black), GH (red), or GHv (green). N = 12–14 per group per sex for all measures. Two-way ANOVA was performed with drug, sex, and interaction (inter.); p-values indicated to the right of each figure. Tukey’s post hoc analyses were performed. * P < 0.05 indicates a significant difference.

Table 1.

Bone mineral density and bone strength as fetermined by uCT of left femora from male and female GH-/- mice treated with saline, GH, or GHv

	Saline	GH	GHv	Saline vs GH	Saline vs GHv	GH vs GHv
Cortical bone	(n = 14)	(n = 14)	(n = 13)
BMD (g/cm3)	1.46 ± 0.04	1.46 ± 0.03	1.47 ± 0.04	P  = 0.987	P  = 0.933	P  = 0.934
MMIp (mm4)	0.08 ± 0.01	0.08 ± 0.00	0.09 ± 0.01	P  = 0.621	P = 0.363	P  = 0.610
BV/TV (%)	33.98 ± 0.80	34.58 ± 0.57	35.33 ± 0.77	P = 0.544	P = 0.236	P = 0.437
T.Ar (mm2)	0.92 ± 0.02	0.93 ± 0.03	0.94 ± 0.03	P = 0.637	P = 0.488	P = 0.815
B.Ar (mm2)	0.31 ± 0.01	0.32 ± 0.01	0.33 ± 0.01	P = 0.524	P = 0.257	P = 0.453
Cs.Th (mm)	0.09 ± 0.00	0.10 ± 0.00	0.10 ± 0.00	P = 0.438	P = 0.273	P = 0.558
Ma.Ar (mm2)	0.60 ± 0.02	0.61 ± 0.02	0.61 ± 0.02	P = 0.779	P = 0.835	P = 0.949
Trabecular bone	(n = 13)	(n = 13)	(n = 12)
BMD (g/cm3)	0.17 ± 0.02	0.19 ± 0.01	0.18 ± 0.03	P  = 0.464	P  = 0.686	P  = 0.749
BV/TV (%)	4.82 ± 0.57	5.25 ± 0.45	5.25 ± 0.58	P = 0.558	P = 0.604	P = 0.996
Tb.Th (mm)	0.04 ± 0.00	0.04 ± 0.00	0.04 ± 0.00	P = 0.790	P = 0.730	P = 0.489
Tb.Sp (mm)	0.46 ± 0.02	0.43 ± 0.02	0.45 ± 0.02	P = 0.236	P = 0.582	P = 0.511
Tb.N (1/mm)	1.16 ± 0.11	1.30 ± 0.08	1.24 ± 0.10	P = 0.341	P = 0.608	P = 0.684

For micro-CT bone measures, no drug effect, sex effect, or interaction effect was observed by 2-way ANOVA (n = 6–7 per group per sex); therefore, male and female data were combined to increase numbers (n = 12–14 per group) and 1-way ANOVA was performed with P-values given in the right 3 columns from Tukey’s post hoc analyses.

Abbreviations: B.Ar, bone area; BMD, bone mineral density; BV/TV, bone volume/total volume; Cs.Th, cortical bone thickness; Ma.Ar, marrow area; MMIp, polar moment of inertia (a measure of bone strength); T.Ar, total cross sectional area; Tb.N, trabecular number; Tb.Sp, trabecular spacing; Tb.Th, trabecular thickness.

Effects of GHv treatments on size of various tissues

In addition to longitudinal bone growth, GH plays an important role in the regulation of tissue/organ size with the removal of GH action, resulting in decreased size of most organs (30). Accordingly, we measured the masses of various tissues to determine the effects of GH and GHv on organ growth. In most nonadipose tissues, organ size was either significantly increased or there was a strong trend toward an increased size following treatment with GH or GHv (Fig. 3). Specifically, gastrocnemius (Fig. 3A), quadriceps (Fig. 3B), and spleen (Fig. 3G) masses were significantly increased in both sexes, with both GH and GHv treatments compared to saline-treated controls. Heart (Fig. 3C), lung (Fig. 3D), and kidney (Fig. 3F) followed the same trend but did not reach significance in one or both sexes. Liver mass was the lone outlier, as livers from GH treated mice were enlarged to a significantly greater extent than livers from GHv-treated mice (Fig. 3E). No effect of either treatment was observed in adipose tissue depots (Fig. 3I-L), except in subcutaneous white adipose tissue (WAT) (Fig. 3H) where both GH and GHv treatment resulted in decreased depot mass in males.

Figure 3.

Effect of GH and GHv on tissue mass. Tissue masses are provided for gastrocnemius muscle (A), quadriceps muscle (B), heart (C), lung (D), liver (E), kidney (F), spleen (G), subcutaneous WAT (H), retroperitoneal WAT (I), mesenteric WAT (J), perigonadal WAT (K), and brown adipose tissue (L). Samples were collected from 5-week-old mice following 6 days of treatment with saline (black), GH (red), or GHv (green). N = 12–14 per group per sex for all measures. Two-way ANOVA was performed with drug, sex, and interaction (inter.); p-values indicated to the right of each figure. Tukey’s post hoc analyses were performed. * P < 0.05 indicates a significant difference.

GHv has less diabetogenic activity compared to GH

The diabetogenic activity of GH has been demonstrated in multiple animals (1, 2) including humans (3, 4, 9). To compare the diabetogenic activity of GH and GHv, we measured serum insulin and performed ITTs in mice after 6 days of treatment. As expected, serum insulin and C-peptide levels were significantly increased (Fig. 4A and 4B) following GH treatment in both sexes; however, neither serum insulin or C-peptide levels were different between GHv treatment and saline-treated controls. Fasting blood glucose levels were significantly increased in males following GH treatment compared to both saline- and GHv-treated mice but were normal in females (Fig. 4C). Insulin tolerance (Fig. 4D–4F) was significantly more impaired in both male and female mice treated with GH compared to saline- or GHv-treated mice. Insulin tolerance with GHv treatment did not differ from saline-treated controls in both sexes.

Figure 4.

GH but not GHv treatment increases circulating insulin levels and decreases insulin sensitivity. Fasting insulin (A), c-peptide (B), and blood glucose (C) levels are shown for male and female GH-/- mice. Samples were collected from 5-week-old mice following 6 days of treatment with saline (black), GH (red), or GHv (green). Male (D) and female (E) insulin tolerance tests and corresponding area-under-the-curve values (F) are shown. N = 12–14 per group per sex for all measures. Two-way ANOVA was performed with drug, sex, and interaction (inter.); p-values indicated to the right of each figure. Tukey’s post hoc analyses were performed. * P < 0.05 indicates a significant difference.

Effects of GHv on TSH, ACTH, FSH, and prolactin in GH-/- mice

Growth hormone can alter other pituitary hormone levels; however, data in the literature are conflicting and no studies have been conducted to determine how GHv affect circulating levels of these hormones. Serum prolactin levels (Fig. 5A) were significantly decreased with GH treatment but not with GHv treatment compared to saline controls in both sexes. Serum prolactin levels did not differ between GHv and controls in males and females. For all other pituitary hormones measured (TSH, ACTH, and FSH), GHv treatment produced a similar response as GH treatment. Specifically, TSH and ACTH levels (Fig. 5B and 5C) were decreased with GH and GHv treatments in both sexes compared to saline-treated controls. FSH levels (Fig. 5D) were increased with GH and GHv treatments compared to saline-treated controls.

Figure 5.

GH but not GHv treatment decreases serum prolactin concentration. Circulating levels of prolactin (A), TSH, (B), ACTH (C), and FSH (D) are shown for male and female GH-/- mice. Samples were collected from 5-week-old mice following 6 days of treatment with saline (black), GH (red), or GHv (green). N = 12–14 per group per sex for all measures. Two-way ANOVA was performed with drug, sex, and interaction (inter.); p-values indicated to the right of each figure. Tukey’s post hoc analyses were performed. * P < 0.05 indicates a significant difference.

GHv has decreased cell proliferation activity compared to GH in three human cancer lines with high PRLR

Since GHv has been shown to have diminished PRLR binding in vitro (11, 13) and because our in vivo study demonstrated that GHv fails to suppress endogenous prolactin levels in GH-/- mice, we tested the hypothesis that GHv will have decreased proliferative effect on human cancers with higher PRLR levels compared to GH. To test this hypothesis, we chose a panel of 6 different human cancer cell lines with varying levels of PRLR and GHR. The PRLR/GHR ratios were calculated using RT-qPCR and varied from high to low in the following order: T84 (colon), T47D (breast), MDA-MB-231 (breast), A498 (kidney), H1299 (lung), and SK-MEL-28 (melanoma) (Fig. 6A and 6B). A 48-hour treatment with equimolar proportions of GH or GHv presented a distinct ability of GH to significantly increase cell proliferation in all 6 human cancer cell lines (Fig. 6C–6I). In contrast to GH, GHv treatment preferentially increased cell proliferation in the cell lines with lower PRLR/GHR ratios (A498, H1299, SK-MEL-28, MALME-3M) but not in the cell lines with higher PRLR/GHR ratios (T84, T47D, MDA-MB-231).

Figure 6.

GHv has reduced cell proliferative activity in human PRLR-positive cancer cell lines compared to GH. A panel of 6 human cancer cell were treated with increasing doses (0, 0.6, 2.5, 10 nM) of GH (red) or GHv (green). Cell growth after 48 hours was determined by comparing cell viability of treated cells to that of nontreated cells using PrestoBlue assay (Invitrogen). A: GHR and PRLR Ct-values and a ratio of 2^-dCT from qPCR. Note: a lower the Ct-value = higher expression. B: PRLR/GHR ratios of the 6 human cancer cell lines. Cell proliferation results are normalized based on treatment with saline (% increase relative to control) and shown for the following human cancer cell lines: (C) T84 (colon cancer), (D) T47D (breast cancer), (E) MDA-MB-231 (breast cancer), (F) A498 (renal cancer), (G) H1299 (lung cancer), and (H) SK-MEL-28 (melanoma). Two-way ANOVA was performed with drug, sex, and interaction (inter.); p-values indicated to the right of each figure. Tukey’s post hoc analyses were performed; * P < 0.05 indicates a significant difference in cell proliferation compared to nontreated control cells and # indicates a significant difference (P < 0.05) in cell proliferation between GH and GHv treatment.

Discussion

By using a large number of GH-deficient mice of both sexes, we were able to robustly demonstrate that GHv had many similar activities as GH with 2 important exceptions. Specifically, similar to that of GH treatment, GHv treatment increased body weight, lean body mass, fluid mass, serum IGF-1, body length, femur length, and FSH compared to saline-treated controls. GHv treatment also decreased body fat mass, TSH, and ACTH similar to GH treatment. However, GHv-treated mice had dramatically improved insulin sensitivity compared to GH-treated mice, and GHv treatment resulted in a significantly decreased proliferative response in 3 separate PRLR-positive human cancer cell lines compared to GH. Thus, our findings demonstrate that GHv can indeed stimulate IGF-1 and subsequent longitudinal body growth and may have clinical relevancy for GH replacement therapy. Furthermore, the 2 activities that distinguish GHv from GH (less diabetogenic action and less ability to promote proliferation of PRLR-positive cancers) may represent improvements to current GH therapies, especially for individuals at risk for type 2 diabetes or select cancers.

To determine if GHv can stimulate IGF-1 and subsequent longitudinal growth, we used young (4-week-old) GH-deficient mice, which have previously been shown to respond to GH treatment and produce IGF-1 (16). Our study demonstrates that GHv treatment clearly increases IGF-1, femur length, and body length compared to saline-treated controls. These data are inconsistent with data from Vickers et al, as GHv did not increase IGF-1 or body length, and bone length was not measured (13). We suspect that these differences are due in part to the numbers of animals used, age of treatment, and the animal model used in each study. In the study by Vickers, a moderate number of 7- to 8-week-old male rats were used with N = 6 per group to observe changes in body length after 1 week of treatment (13). Furthermore, while body length was not significantly increased in their study, Vickers et al did report a trend toward an increased nose-to-tail length (P = 0.06) with GHv treatment, suggesting that an increase in animal numbers may have yielded a significant result (13). Regarding the animal model used, wild-type laboratory rodents have been described by Bielohuby et al to have a limited IGF-1 response to exogenous GH treatment despite changes to glucose homeostasis, body composition, and organ mass (14, 15, 17). The authors propose that IGF-1 release may be near maximal in wild-type rodents, making further increases by exogenous GH treatments difficult to achieve (14). Therefore, using a non-GH deficient rodent model likely made it difficult to observe changes in IGF-1 and subsequent longitudinal growth. In order to overcome some of these challenges, we used (1) a larger number of animals n = 12–14 per group per sex, (2) both sexes, and (3) a newly developed GH-deficient mouse line (16). Accordingly, our results indicated that GHv is fully capable of stimulating IGF-1 and subsequent longitudinal bone growth in a GH-deficient model, and our study demonstrates that GHv treatment worked as well as GH treatment in both sexes. However, the lack of change in bone parameters (such as bone mineral density) with either treatment suggests longer treatment periods may be needed to see these changes.

Our data indicate that GHv alters body composition similar to GH, but unlike GH, lacks diabetogenic activity in young GH-deficient mice. We found that GHv treatment significantly increased body mass and fluid mass while decreasing fat mass similar to GH in both male and female mice. Data in male mice are in agreement with data from Vickers et al, as GHv significantly reduced body fat as well as the fat-to-lean ratio in high-fat fed rats similar to GH; however, fluid mass was not measured and lean mass was only represented in the fat-to-lean ratio (13). While our data showed that GHv had the full ability to alter body composition similar to GH (and stimulate IGF-1 and growth, as discussed above), in contrast, GHv lacked the diabetogenic activity of GH. More specifically, GHv failed to cause hyperinsulinemia, and mice treated with GHv were significantly more insulin sensitive compared to mice treated with GH. These findings also agree with data from Vickers (13), as GHv treatment in high-fat fed rats resulted in significantly reduced insulin and C-peptide levels compared to GH treatment; however, insulin sensitivity was not measured and insulin levels in GHv-treated rats were even lower than saline-treated controls and only males were evaluated. Thus, taken together, 2 separate laboratories using 2 separate species and differing methods have demonstrated that GHv has significantly diminished diabetogenic activity while retaining its full ability to enhance body composition. Collectively, these data provide strong evidence for evaluating the use of GHv in humans with growth disorders or with obesity.

GHv altered circulating levels of ACTH, TSH, and FSH similar to GH, but in contrast to GH, did not reduce prolactin in young GH-/- mice. In our study, both GH and GHv treatments increased FSH compared to saline-treated controls. Since basal and stimulated plasma FSH levels are attenuated in GH-immunoneutralized, GH-deficient, and GH-resistant animals (31–35), and exogenous GH has been shown to increases release of FSH from rodent pituitary glands (36, 37), the increase in FSH following GH treatment in GH-/- mice is expected. However, this relationship between GH and FSH is not always consistent since GH exerts negligible or inhibitory effects on FSH secretion in other studies (38–41). We also observed that GH and GHv treatments reduced TSH and ACTH. Evidence for GH treatment reducing TSH has been observed in both children and adults treated with GH (42–47) and daily administration of GH for 4 to 6 days in humans inhibits ACTH release (48). While alterations in these hormones are reported following GH administration, it should be noted that there are wide inconsistencies in the literature about these perturbations (as reviewed in 49, 50).

The only pituitary-derived hormone that distinguished GH versus GHv treatment was prolactin, with GH treatment causing reduced circulating prolactin levels, while GHv did not. This result may be due to the extreme difference in lactogenic activity between the 2 GHs. In 1990, Cunningham and his colleagues demonstrated that human GH binds to PRLR with high affinity (Kd of hGH binding to PRLR in the presence of zinc is 0.033 mM, 78x better than PRL for its own receptor at 2.6 mM (51)). In contrast, Solomon (11) and Vickers (13) demonstrated that GHv has no lactogenic response. Because of this difference, mice treated with high-dose hGH should trigger prolactin’s negative feedback loop (52) where stimulation of PRLR in the hypothalamus increases dopamine, which, in turn, inhibits prolactin release from the pituitary. Since GHv does not have lactogenic activity, treatment with GHv should not affect prolactin’s negative feedback loop. In agreement with this observation, human GH transgenic mice repeatedly exhibit decreased circulating prolactin levels (36, 53–56). In contrast to hGH, mice with high bovine or mouse GH, neither of which is reported to have lactogenic activity, do not exhibit reduced prolactin; therefore, this prolactin-lowering effect in mice appears to be specific to hGH (31, 36, 57, 58).

Two separate laboratories have previously established that GHv has no lactogenic activity (no detectable affinity for the PRLR) (11, 13). We feel that this finding is clinically relevant since prolactin is implicated in the progression of certain types of cancers including but not limited to breast (59, 60), prostate (59, 61), colon (62), and ovarian (63, 64) cancer. Growth hormone is also implicated with the progression of many of these same cancers (65–71). This is not surprising since human GH binds to both the GHR and PRLR; thus, the promiscuous binding of GH to PRLR has been theorized to be at least partially responsible for GH’s ability to stimulate these cancers (65, 67, 68). Because GHv has diminished lactogenic activity, we chose to test the ability of GHv to stimulate proliferation of PRLR-positive cancers. Of the 6 human cancer cell lines we tested, 3 showed a significant reduction in proliferation when treated with GHv compared to GH. Not surprisingly, these 3 lines T84 (a human colon cancer cell line), T47D (a human breast cancer cell line), and MDA-MB-231 (a human breast cancer cell line) had the highest levels of PRLR mRNA expression and the highest PRLR/GHR ratios. Thus, our data agree with the theory that the lactogenic activity of GH may be partially responsible for stimulating proliferation of certain PRLR-positive cancers, such as colon and breast. Importantly, our data also suggests that GHv may represent improvements to current GH therapies, especially for individuals at risk for PRLR-positive cancers.

When translating the potential effects of GHv from mice to humans, it is important to note that several laboratories have previously reported that the 20K isoform of a different gene, GH-N, has reduced diabetogenic activity in dogs (72) and rats (73–75). However, these reports are controversial since other studies report that 20K GH-N has diabetogenic activity in these same species (76, 77). Furthermore, an evaluation of 20K GH-N in humans failed to detect a difference in diabetogenic activity between the 20K versions of GHv and GH-N (78). However, the diabetogenic activity of GH is a transient affect, and thus measures of insulin sensitivity should be performed shortly after GH injection when serum GH levels have not returned to baseline levels (79, 80). Unfortunately, Hayakawa et al tested subjects immediately prior to GH treatment, making the detection of diabetogenic activity unlikely to occur (79, 80). Furthermore, this study did not utilize a 22K GH-N positive control group (78); thus, direct comparison of 20K GH-N to 22K GH-N has yet to be conducted in humans. Regardless of these results, it is important to remember that GH-N and GH-V are 2 distinct genes and their 20K isoforms differ by 12 amino acids from one another. This difference in amino acid identity may be substantial, since it has previously been shown that it only takes a single amino acid change in human GH to completely change its activity (81, 82). Furthermore, binding studies have already shown opposing activities between these 2 hormones, as GH-N (both 22K and to a lesser extent 20K, ~ 4.4-fold less) can bind to human, rat, and rabbit PRLR (11); in contrast, GHv has no detectable binding affinity for PRLR in all species tested including human PRLR (11, 13). Thus, there is little reason to believe that this activity (or lack thereof) will not translate to humans. Taken together, at minimum, GHv represents a GH therapeutic that is unable to stimulate PRLR-mediated growth in PRLR-positive human cancers and, at maximum, GHv may have additional, difficult to detect benefits to glucose homeostasis compared to conventional GH therapy.

In summary, we demonstrate that GHv (20K GH-V) is a potent stimulator of body growth, as indicated by significant increases to IGF-1, femur length, body length, body weight, and lean body mass, and also reduces body fat mass similar to GH treatment in a mouse model of GH deficiency. Our findings are the first to demonstrate that GHv can indeed stimulate IGF-1 and subsequent longitudinal body growth, and may have clinical relevancy for GH replacement therapies where body growth is the desired outcome. Moreover, GHv-treated mice had reduced insulin levels and dramatically improved insulin sensitivity compared to GH-treated mice, which is consistent with findings from Vickers (13) in high-fat fed rats and suggests that GHv has diminished diabetogenic activity. Importantly, GHv appears to be efficacious in both sexes. Finally, we demonstrate that GHv treatment has significantly decreased proliferative response in 3 separate PRLR-positive human cancer cell lines compared to GH. Thus, GHv may represent improvements to current GH therapies, especially for individuals at risk for metabolic syndrome or PRLR-positive cancers.

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.