Materials

Unless otherwise stated, chemicals and reagents were purchased from either Merck Darmstadt, Germany) or Thermo Fisher Scientific (Waltham, MA, USA). Hydrochloric acid (HCl; 32% w/v AnalaR) was supplied by VWR International (Radnor, PA, USA). Protein Lo-Bind Eppendorf tubes (1.5 ml) were supplied by Eppendorf (AG Hamburg, Germany). Sodium chloride (NaCl), and tris(hydroxymethyl)aminomethane (Tris) were purchased from Astral Scientific, (Sydney, NSW, Australia). Fetal bovine serum was purchased from Interpath (Heidelberg West, Vic, Australia). All peptides were purchased from GenScript (Piscataway, NJ, USA) with C-terminal amidation and more than 98% purity. All cell lines were purchased from ATCC (Manassas, Virginia, USA) except the mECap18 cells that were obtained as a generous gift from Dr Petra Sipilä (Institute of Biomedicine, Department of Physiology, University of Turku, Turku, Finland). Bioss rabbit polyclonal anti-LH receptor antibody (bs-6431R) and donkey serum were purchased from Sapphire Bioscience (Redfern, NSW, Australia). Goat polyclonal anti-8-hydroxydeoxyguanosine (8-OHdG) was purchased from Millipore (Temecula, CA, USA). Agarose (low melting point) was obtained from Applichem GMBH (64291 Darmstadt, Germany) and ethanol purchased from POCD Scientific (North Rocks, NSW, Australia). DNA/RNA Damage Antibody (15A3) was obtained from Novus Biologicals (Littleton, CO, USA). Maleimide-vcMMAE (Auristatin) was purchased from MedChemExpress (Monmouth Junction, NJ, USA).

Design of LH receptor targeting peptide

To enable specific targeting of Leydig cells, a peptide was designed that possessed high binding capability and specificity as well as the in vivo stability required for the peptide to reach its target. We have previously developed a peptide based on the residues of the β-subunit of FSH that specifically targets Sertoli cells [22]. Here, we employed the same design principles to design a peptide corresponding to residues occupying the equivalent “seatbelt” region of the mouse LH β-subunit; a region known to be important for gonadotrophin receptor binding [24–27]. An important modification of the native sequence was that all cysteine residues, except those that occurred on the end of the molecule, were substituted with serine residues. This modification most closely mimics the native 3-D structure of the peptide, yet reduces the propensity for peptide cyclization to occur. Moreover, removal of internal reactive thiol groups ensured a single site was available for maleimide-thiol mediated coupling of the peptide to the desired payload and thus avoided the potential for mixed end products to be formed. Finally, the C-terminal was amidated to inhibit protease degradation of the peptide in vivo.

Physicochemical properties and structure of LH peptides

A comparison was made of the physicochemical properties of LH synthetic peptides using data obtained from PepCalc (PepCalc.com, 2015 Innovagen AB). Furthermore, Swiss-Prot Deep Viewer software was used to map each of the peptides onto the core structure of LHβ Swiss-Model O09108 (LSHB_MOUSE) Mus musculus (Mouse) Lutropin subunit beta Template: 1hcn.2.B "HUMAN CHORIONIC GONADOTROPIN" P0DN86SMTL Version: 2019-12-06, Seq Identity: 63.96%, Seq Similarity: 0.52.

Cell culture

All cell lines were maintained in a humidified atmosphere at 37°C and 5% CO₂. MLTC1 Leydig cells were cultured in growth medium composed of RPMI 1640 medium supplemented with 10% (v/v) fetal bovine serum, and 100 U/mL penicillin-streptomycin. TM3 Leydig cells were cultured in growth medium composed of F12/Dulbecco’s modified Eagle’s medium, 2.5 mM L-glutamine, 0.5 mM sodium pyruvate, 15 mM HEPES, 2.5% of a mixture of 92.5% horse serum and 5% fetal bovine serum, and 100 U/mL penicillin-streptomycin. MLTC1 and TM3 cells were sub-cultured at a ratio of 1:5 into cell culture flasks. Immortalized mouse proximal caput epididymal epithelial (mECap18) cultures [28] were used as a nonspecific cell control for investigating the specificity of LH peptide interactions and were maintained in growth medium composed of DMEM supplemented with 10% (v/v) fetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin-streptomycin, 1 mM sodium pyruvate and 50 nM 5-α-androstan-17β-ol-3-one and sub-cultured at a ratio of 1:10 into culture flasks. Cells were grown to 80% confluency before being harvested using 0.5% trypsin-EDTA, quenched, centrifuged at 300 × g for 5 min, and then washed in warmed growth medium. After centrifugation, the cells were resuspended in warm medium and subjected to incubation with peptide- fluorescein isothiocyanate (FITC) conjugates.

Animals and animal care, injections and monitoring

In this study, adult (>8 week old) male Swiss (CD1) mice were obtained from a breeding colony at the University of Newcastle Central Animal House and housed under a conventional controlled light and temperature regimen (12-h light: 12-h dark cycle, 21–22°C). All procedures involving mice were conducted in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health. Additionally, mice were handled, monitored and euthanized with the approval of the University of Newcastle’s Animal Care and Ethics Committee (approval number A-2014-417) in accordance with NSW Animal Research Act 1998, NSW Animal Research Regulation 2010 and the Australian Code for the Care and Use of Animals for Scientific Purposes 8th Ed.

Validation of the MLTC1 Leydig cell model

To ascertain the binding capability of synthetic LH peptides, we required a Leydig cell line that expressed the LH receptor. The two mouse Leydig cell lines, TM3 and MLTC1, are both expected to express the LH receptor. RNA was extracted from testis, MLTC1 Leydig cells, TM3 Leydig cells and the mouse epididymal epithelial (mECap18) cells, with the latter serving as a negative control. To determine which of the two Leydig cell lines had the highest relative LHr expression quantitative PCR (qPCR) was used to assess expression relative to the house-keeping gene, GAPDH. Subsequently, LHr protein expression was assessed in cultured cells using anti-LHr antibodies.

RNA extraction, reverse transcription and quantitative PCR (qPCR)

Total RNA isolation was performed using two rounds of a modified acid guanidinium thiocyanate–phenol–chloroform protocol [29] followed by isopropanol precipitation. Total RNA was DNase treated prior to reverse transcription to remove genomic DNA. Reverse transcription to cDNA was conducted using M-MLV Reverse Transcriptase (Promega, M1701) following the manufacturer’s instructions with subsequent confirmation of transcription using cyclophilin qPCR (cyclophilin primers (5’-CGTCTCCTTCGAGCTGTTT-3’; 5’-ACCCTGGCACATGAATCCT-3’); T_m 61˚C).

Quantitative PCR: Expression of LHr relative to, the housekeeping gene, GAPDH, was conducted as follows: Taq qPCR was performed using SYBR Green GoTaq qPCR master mix (Promega) according to manufacturer’s instructions on a LightCycler 96 SW 1.0 (Roche Diagnostics GmbH, Mannheim, Germany) thermocycler. Primer sequences for the LH receptor and GAPDH have been supplied as supplementary data (S1 Fig). 20 μl reactions were performed on cDNA equivalent to 100 ng of total RNA and conducted for 45 amplification cycles. SYBR Green fluorescence was measured after the extension step at the end of each amplification cycle and quantified using LightCycler Analysis Software (Roche). For each sample, replicates omitting the reverse transcription step were undertaken as negative control. Each PCR was performed on at least 3 separate cell or tissue isolations, of which a representative average is shown. C_t values less than 30 were assumed to be primer dimers and were treated as negative for expression. Relative expression was calculated using the ΔΔC_t method.

Validation of MLTC1 Leydig cells

Immunocytochemistry was conducted to confirm expression of the LH receptor protein in MLTC1 Leydig cells using mECap18 cells as a negative control. Cells were grown to 80% confluence on sterile poly-L-lysine-coated round coverslips within the wells of 24-well culture plates. Cells were fixed in situ with 4% paraformaldehyde (PFA) for 10 min at room temperature (RT). After rinsing with buffer 1 [phosphate buffered saline (PBS) with 3% BSA and 0.015% Triton X-100], cells were blocked for 1 h at RT with buffer 2 (PBS with 10% goat serum, 3% BSA and 0.015% Triton X-100) and then incubated overnight at 4°C with Bioss rabbit polyclonal anti-LHr (diluted to 4 μg/ml in buffer 1). Cells were then re-blocked in buffer 2 for 15 min at RT before incubating for 1 h at RT in 1/200 of goat anti-rabbit AF488 in buffer 2. Finally, cells were washed with buffer 3 (PBS plus 0.015% Triton X-100) then counterstained using a Far Red nuclear stain (diluted 1/2000 in PBS) for 30 min at RT. After washing in PBS, coverslips were removed from the wells and inverted onto a drop of Mowiol on glass microscope slides then sealed using clear nail varnish before cells were imaged using fluorescence microscopy.

Assessment of in vitro peptide binding capability and specificity

MLTC1 cells and mECap18 cells were used to assess the relative binding capabilities and the specificity of targeting of the LH peptides. All peptides (Table 1) were purchased with an N-terminal spacer, 1,6-aminohexanoic acid (Ahx), attached to FITC and dissolved in DMSO to form a 2.5 mM stock solution. Assessment of binding capability was conducted as previously described [22]. Briefly: Cells were grown to 80% confluency before harvesting and treatment with the LH-FITC in concentrations ranging from 0 to 50 μM. The cells were incubated at 37°C for 1 h. Cells were then stained with Live/dead Live/dead Far Red stain incubated at 37°C for 20 min before washing and resuspension in HBSS. FITC staining and was then analyzed using fluorescence assisted cell sorting (FACS) as previously described [22]. The median fluorescence intensity was recorded for each treatment group and the shift in median fluorescence intensity was then calculated by subtracting the median fluorescence intensity of the vehicle (0 μM) from the median fluorescence intensity of each dose.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Amino acid sequences and properties of the unmodified LH receptor binding peptides.

https://doi.org/10.1371/journal.pone.0292198.t001

Assessment of in vivo targeting and off-target binding by LH peptides

All LH peptides were assessed in vivo to determine the efficacy of targeting the testes and the extent of off-target binding. A total of 300 μL of 1 mM peptide-FITC or the vehicle (30% Kolliphor/PBS) was injected intraperitoneally (IP) into adult male Swiss CD1 mice, with three males being randomly assigned to each treatment group. Mice were euthanized 8 h post-injection and tissues (adrenals, brain, epididymides, kidneys, liver, lungs, pituitary, seminal vesicles, spleen, and testes) were collected, weighed and immediately imaged. Tissues were imaged in cold PBS with a Leica MZFLIII microscope with an epifluorescent green filter. All images included the vehicle control to enable comparison of median fluorescence intensity of the treatments relative to the control using ImageJ software [ImageJ version 1.52a; Wayne Rasband, National Institutes of Health, USA. http://imagej.nih.gov.ij, Java 1.8.0_112 (64-bit)].

Conjugation of Auristatin to LH peptide

LHa peptide (Table 1) was obtained from GenScript with a γ-aminobutanoate-mercaptopropionic acid linker (hereinafter referred to as LH2 peptide) and was attached, via the N-terminal thiol, to maleimide-Auristatin (in the form of maleimide-valine-citrulline- para-aminobenzyloxycarbonyl -monomethyl-Auristatin E) as follows. LH2 peptide (MW 1333) was dissolved in 10 mM PBS, pH 7, at a concentration of 8.5 mg/250 μL. The peptide was then reduced, using immobilized TCEP (tris(2-carboxyethyl)phosphine) disulphide reducing gel, the solution pH adjusted to 7, and then deoxygenated. Meanwhile, 1.2 molar equivalents of maleimide-Auristatin (MW 1316) were dissolved in 53% dimethyl formamide/10 mM PBS, pH 7, at 10mg/3 mL and placed in a reaction vessel. The maleimide-Auristatin solution was then deoxygenated, and the reduced peptide solution was added dropwise, using a syringe, whilst stirring and under positive pressure with nitrogen gas. The reaction was then stirred overnight at RT. The success of the conjugation reaction was confirmed by mass spectrometry, using a Bruker Ultraflextreme MALDI-TOF/TOF (mode: reflectron positive; matrix: α-cyano-hydroxy-cinnamic acid). The solvents were removed from the reaction by freeze drying under vacuum. The dried reaction mix was then reconstituted and purified using reverse phase HPLC. Samples were eluted from the Waters 250 × 10.0 mm Jupiter 4u Proteo 90A column using a gradient method with Solvent A (0.1% v/v formic acid in H₂O) and Solvent B (90% acetonitrile with 0.1% v/v formic acid) with a flow rate of 2 mL/min over 37 min followed by 2 min at 100% Solvent B before re-equilibration for 5 min. The eluted species were detected by their absorbance at 220 nm. Peak fractions were collected and LH2-Auristatin identity and purity was verified using Orbitrap liquid chromatography/mass spectrometry before the product was freeze-dried under vacuum and stored as a powder at -20°C.

Conjugation of FSH2 peptide with menadione

FSH peptides were attached, via the N-terminal thiol, to menadione as previously described [22]. Briefly, deoxygenated aqueous solution of reduced FSH2 peptide was added dropwise to 10 molar equivalents of menadione in DMSO in a sealed reaction vessel, heated in a water bath to 55°C, whilst stirring with a magnetic bar and under positive pressure of nitrogen gas. After peptide addition to the menadione suspension, stirring of the reaction was continued at 55°C for 2 h. Mass spectrometry was used to confirm the success of the conjugation reaction, using a Bruker Ultraflextreme MALDI-TOF/TOF as above. The reaction product was dried and then purified using reverse phase HPLC. Peak fractions were collected and peptide-quinone identity and purity was verified using MALDI-TOF/TOF mass spectrometry, as above, before the product was freeze dried under vacuum and stored as a powder at -20°C.

In vivo pilot study FSH2Menadione with LH2Auristatin

Four treatment groups of ten 55-day-old male Swiss male CD1 mice were injected intraperitoneally. Treatments included: (1) 00 μL vehicle control (30% Solutol (Kolliphor)/PBS), (2) 300 μL/30 g 14.5 mM FSH2Menadione, (3) 100 μL/30g 425 μM LH2Aur, or (4) 300 μL/30 g 14.5 mM FSH2Menadione plus 100 μL/30g 425 μM LH2Aur (consecutive injections, 16 h apart). The dose of 1 mg/kg of conjugated Auristatin was determined with reference to the work by Doronina et al. [30], who determined the maximum tolerated dose (MTD) of conjugated Auristatin for mice is 1.1 mg/kg. Solubility of the FSH2Menadione was limited to 14.5 mM (25 mg/kg) in the biocompatible reagents (Kolliphor/PBS) delivering a dose much lower than menadione’s MTD of 830 mg/kg. Six weeks post injection, five randomly selected mice from each treatment group were mated with two untreated females (10–12 weeks old) each for three consecutive weeks. Females were monitored for mating plugs each day and on the thirteenth day post-mating plug, or if weight gain indicated approximately 13 days of pregnancy, females were euthanized weighed, embryos and resorption moles were counted, ovaries were collected, weighed, and corpora lutea counted. After the mating period, at ten weeks post-injection, all males were euthanized, weighed and blood collected by cardiac puncture. Blood samples from each animal were decanted into EDTA-coated tubes and plasma was separated (10 min, 400 g). Plasma was stored at -80°C for later measurement of serum FSH levels, for which a AVIVA FSH ELISA Kit was purchased from Sapphire Bioscience (Redfern, NSW, Australia) and conducted as per manufacturer’s instructions.

Tissues (seminal vesicles, epididymides, adrenals, kidneys, liver, brain, heart, spleen and testes) were collected and weighed. A sample of each tissue was fixed in Bouin’s solution (75 mL picric acid, 40% aqueous solution; 25 mL formalin, 40% aqueous solution; and 5 mL glacial acetic acid) for around 40 min per mg tissue. One testis from each male was fixed and the other snap frozen in liquid nitrogen and stored at -80°C. After fixing, testes were washed with 70% ethanol until the solution was clear and then processed for paraffin embedding, sectioning (5 μm thickness) and haematoxylin and eosin staining (H and E). H and E-stained sections were used to evaluate cross-sectional tubule areas as described below. To evaluate the impact on Sertoli cells numbers, an anti-SOX9 antibody was used to stain sections, while an ApopTag Kit (Merck) was used to detect apoptotic cells in accordance with the manufacturer’s instructions.

Immediately after dissection, epididymal spermatozoa were also collected for analysis of motility, vitality, DNA strand breaks (HALO), DNA fragmentation (SCSA), the presence of reactive oxygen species (MSR and DHE) and oxidative DNA adducts (anti-8-OH-dG). For these assays, three incisions were made in the cauda epididymides with a surgical blade before placing the tissue in prewarmed Biggers-Whitten-Whittingham (BWW) medium [31] with 5 U/ml penicillin, and 5 mg/mL streptomycin, pH 7.4, at 37°C. Spermatozoa were allowed to swim-out for 10 min before being collected into clean tubes. Sperm motility was assessed using a HTM-IVOS II Computer Assisted Sperm Analysis and sperm vitality was assessed using the eosin exclusion test [32] and viewed using a light microscope at 40× magnification. At least 100 spermatozoa per sample were analysed to determine percentage vitality of each sample. Spermatozoa were stained for FACS analysis of mitochondrial and cytosolic superoxide generation using MSR and DHE staining [33, 34]. Additionally, sperm smears were made for staining with Rapid Diff (Pathtech, Preston, Vic, Australia), as per manufacturer’s instructions, to evaluate sperm morphology. Analyses of DNA strand breaks (HALO), and detection of the oxidative DNA adducts (anti-8-OH-dG) were conducted as previously described [22]. At least 100 sperm/sample (10,000 sperm/sample for FACS) were assessed for each assay and, where possible, all counts were conducted blind. The remaining spermatozoa were aliquoted into tubes containing 10⁶ sperm/mL, snap frozen in liquid nitrogen and stored at -80°C until use. Thawed sperm cells were later assessed using FACS for DNA fragmentation using a sperm chromatin structure assay (SCSA) as previously described [35, 36]. All FACS data were acquired and analyzed using BD FACSDiva software (BD Biosciences), with a total of 10,000 events collected per sample. Only results for viable cells are reported for all FACS assays.

Cross-sectional area of tubules

To ascertain average seminiferous tubule area, six H and E-stained testes sections were randomly chosen from each treatment group. Sections were imaged using a 10 × objective under bright field microscopy and then images (4080 × 3072 pixels) were imported into ImageJ software. At least 50 tubules from each testis section were outlined and the area in pixels/tubule calculated.

SOX9 staining to ascertain Sertoli cell number

Testes sections (5μm thick) were dewaxed using xylene and then rehydrated in an ethanol gradient. Heat-assisted epitope retrieval was conducted in Tris-EDTA, pH 9. Slides were then incubated in 0.3% hydrogen peroxide in Tris buffered saline (TBS) for 30 min at RT. After washing in TBS (3 × 5 min), the sections were blocked using 5% normal goat serum (NGS) for 30 min at RT in a humidified chamber. The sections were then incubated with rabbit anti-SOX9 antibodies (Merck: AB5535) in NGS (1:500) overnight at 4°C. Slides were washed in TBS (3 x 5 min) prior to incubation with biotinylated goat anti-rabbit antibodies (Abcam, ab6720) in NGS (1:200) for 30 min in a humidified chamber at RT. Slides were washed in TBS (3 × 5 min) then incubated with streptavidin-HRP (Abacus SA-5004) in TBS (1:1000) for 30 min at RT before being washed again in TBS (3 × 5 min). Sections were then treated with 3,3’-diaminobenzidine (DAB) for up to 5 min or until sufficient colour development. Slides were washed with TBS (3 × 5 min) and counterstained with haematoxylin for 8 min before being dehydrated and mounted with Eukitt. Slides were imaged using bright field microscopy and the number of Sertoli cells counted in all tubules across one section from each slide.

Statistical analysis

All experiments were conducted at least three times on individual samples in technical triplicate. All quantitative data was expressed as the mean ± standard error from three or more replicate samples. Statistical analysis of data was conducted using Microsoft Excel (v16.0; Microsoft Corporation, Redmond, WA, USA) as follows: a single factor ANOVA was conducted on the data from all treatments. If the p-value was < 0.05, then pair-wise Student’s t-tests were conducted using two-samples assuming unequal variance if difference between the variances was > 2-fold, otherwise assuming equal variance.

Design of LH receptor binding peptides

A LH receptor binding peptide (hereinafter referred to as LHa) was designed based on equivalent residues to those incorporated into a previously developed FSH peptide mimic [22]. Specifically, the LHa peptide incorporated amino acid residues mapping to LH β-subunit 96–107 (Table 1). An additional three LH peptides were also designed based on sequences that confer specific receptor binding. These were designated L57, L95 and L101 depending on where their sequence mapped within the LH β-subunit. L57 peptide included the amino acid residues LHβ 38–57 [26], L95 peptide the residues in LHβ 81–95 [25] and L101 peptide included the residues LHβ 93–101 [23] (Table 1).

All LH peptides mapped to the surface of the receptor binding regions of LH β-subunit

The position of the peptides mapped to surface of the 3-D structure of the LH β-subunit (Fig 1) confirmed that the new mouse amino acid sequences were located in the βL2 loop and the “seatbelt”, which are the two binding regions on the surface of the β-subunit (Fig 1). Additionally, their position on the surface of the protein means that they are accessible for receptor binding. The positioning of the peptides on the surface of the protein within the recognised binding regions was expected to confer a high likelihood of binding to the LH receptor.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Mapping the peptides to the β- subunit of the LH protein which targets receptors on theca cells in the ovary and Leydig cells of the testis.

LHa rationally designed based on binding residues of β-subunit of FSH protein with FSH receptor [37]. Swiss-Prot Deep Viewer was used to map the peptides (in green) to the surface of Swiss-Model O09108 (LSHB_MOUSE) Mus musculus (Mouse) Lutropin subunit beta Template: 1hcn.2.B "HUMAN CHORIONIC GONADOTROPIN" P0DN86SMTL Version: 2019-12-06, Seq Identity: 63.96%, Seq Similarity: 0.52. The βL2 loop and the seatbelt region are indicated by arrows and are the receptor-binding regions of the β-subunit of the glycoprotein hormones.

https://doi.org/10.1371/journal.pone.0292198.g001

A comparison of the physicochemical properties of the peptides (Table 1) predicted good aqueous solubility for all of the peptides except L57. This may be expected as the ratio of hydrophobic to hydrophilic residues for this peptide was four, considerably higher than that of the other, more soluble, peptides which all had an equivalent ratio of less than one. The predicted charge at neutral pH on LHa and L101 were both around +1. L57 was predicted to have a higher overall charge of +1.9, whereas L95 was predicted to possess a very high +3.9 charge at neutral pH. In recognition that both hydrophobicity and charge can influence the affinity of a peptide for the cell surface (with positive charges increasing the electrostatic attraction for negative cell surfaces), we elected to directly assess the binding specificity of each peptide (see section entitled ‘LH peptides demonstrated variable in vitro binding capability and specificity’).

Q-PCR and immunocytochemistry confirmed MLTC1 Leydig cells as an in vitro model

For this study, as we aimed to target the LH receptor, we required a cell line that displayed strong expression of the LH receptor gene (Lhcgr) to act as an in vitro model. qPCR and an anti-LHr antibody were used to probe for the presence of the LH receptor in both MLTC1 and TM3 Leydig cell lines, as well as in the control epididymal epithelia cell line (mECap18). Using qPCR, we confirmed that of the two available mouse Leydig cell lines, the MLTC1 Leydig cell line expressed the Lhcgr gene (S2 Fig). By contrast, no Lhcgr transcripts were detected in the TM3 cell line or in the negative control mECap18 cell line. These data were validated by immunocytochemistry, which confirmed LH receptor expression within the MLTC1 Leydig cells, but not in mECap18 cells (S2B Fig). Given these data we elected to use MLTC1 cells as the in vitro Leydig cell line model.

LH peptides demonstrated variable in vitro binding capability and specificity

MLTC1 cells and mECap18 cells were used to assess the relative binding capabilities and the specificity of targeting of synthetic LH peptides, revealing that each peptide bound to the MLTC1 Leydig cells in a dose-dependent manner. This was evidenced by a proportional shift in the median fluorescence intensity of the stained MLTC1 cell population across the dose range of applied peptides (Fig 2A). Additionally, each of the peptides displayed suitable specificity with three-fold higher median fluorescence intensity when applied to MLTC1 Leydig cells as compared to the mECap18 control cells (Fig 2B and 2C). The peptide with the greatest shift in median fluorescence and, by inference, the greatest binding capability, was L95 (Fig 2B); which possessed more than double the MLTC1 cell binding affinity than that of the other peptides tested. However, L95-FITC showed a lack of specificity at the highest dose tested, 50 μM, with this peptide exhibiting only 1.5-fold relative binding on MLTC1 cells above that of the background mECap18 cell binding.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Peptide targeting and binding affinity in vitro.

Peptide-FITC in DMSO or DMSO was applied 1 in 100 parts (v/v) to the cells to make final concentrations of 0 to 50 μM for 1 hour at 37°C. A. FACS analysis of binding of FITC-conjugated LH peptides to MLTC1 Leydig cells (n = 3) using shift in median FITC fluorescence intensity in response to increasing concentration. B. Comparison of binding of LH-peptides to MLTC1 Leydig cells (n = 3), and C. mEcap18 cells. FACS analysis of binding of the peptides using shift in median fluorescence intensity (n = 3).

https://doi.org/10.1371/journal.pone.0292198.g002

LH peptides localized to the testes in vivo

Given the promising in vitro results, we sought to verify the homing potential of the LH-FITC conjugates to their testicular target following injection into adult male mice. We failed to detect any significant differences in the weight of mice exposed to either the vehicle or the peptide-FITC treatments eight hours post-injection. Similarly, we did not record any differences in the weight or gross morphology of the representative tissues collected. All four LH peptides exhibited localization within the testes with a clear difference in the fluorescence intensity being detected in the testes of the vehicle-injected controls, compared to that of the peptide-FITC injected males (Fig 3A). This contrasted the findings of the adrenals, brain, lungs, pituitary, or seminal vesicles for which no difference in fluorescence staining was observed between the vehicle control and the treated animal tissues. However, there was evidence of clearance of the peptides through the detoxifying and excretory organs of the liver and kidney, both of which presented with relatively low fluorescence labelling (S3 Fig). Additionally, there appeared to be modest FITC fluorescence associated with the spleen in at least a portion of the treated animals, but particularly within the spleens of the L57-FITC treated animals (S3 Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Confirmation of in vivo targeting by LH peptides.

Representative images of peptide-FITC localised to testes 8 hours post-injection. Male adult mice were injected IP with 300 μl of either 1mM peptide-FITC in 30% Kolliphor/PBS or 30% Kolliphor/PBS with 3 mice in each treatment group. Animals were euthanized 8 hours post-injection and the testes were imaged, using Leica MZFLIII microscope with an epifluorescent green filter, next to the vehicle control. A. Representative images of testes after injection with (from top to bottom) LHa-FITC, L57-FITC, L101-FITC and L95-FITC. B. Image J analysis of mean fluorescence intensity/pixel in testes of peptide-FITC injected males. Images were analysed using ImageJ software (ImageJ 1.52a, Wayne Rasband, National Institutes of Health, USA. http://imagej.nih.gov.ij, Java 1.8.0_112 (64-bit)). Mean fluorescence intensity/pixel was normalised to that of the vehicle control within the same image. (n = 3). Statistical analysis by One-way ANOVA then post hoc using student T-tests.

https://doi.org/10.1371/journal.pone.0292198.g003

Based on these findings, images of the treated testes were imported into ImageJ to compare the mean fluorescence intensity/pixel. The results were expressed as a ratio of the mean fluorescence intensity/pixel of the treated in any image to the vehicle control in the same image (Fig 3A). All four peptides displayed a similar normalized fluorescence intensity/pixel of between 1.45 and 1.6-fold that of the vehicle control (Fig 3B), albeit with differing levels of significance. These results suggest that the LH peptides possess similar abilities to localize to the testes and that they possess similar stabilities in vivo. The presence of the FITC-Ahx tag complexed to the peptide, however, means that important factors that determine the pharmacokinetics, such as the theoretical isoelectric point, the charge and the logP value (a measure of hydrophobicity that increases with increasing hydrophobicity) of the complex differs from that of the unmodified peptide and, most likely, that of the peptide with cargo. Therefore, whilst informative, this method of peptide in vitro and in vivo tracking may not be definitive, and additional criteria were investigated to aid in our selection of an appropriate LHr targeting peptide.

Among these criteria, aqueous solubility was considered an important factor, both in the context of reaction with Auristatin and for use of biocompatible solvents for in vivo delivery. The Auristatin employed in this study has a valine-citrulline linker (S4 Fig) with a terminal maleimide for ease of conjugation and has very poor aqueous solubility with a logP value of 6.04 (https://www.chemsrc.com/en/cas/646502-53-6_1197872.html). Therefore, so as not to compound this lack of solubility, the targeting peptide needed to possess good aqueous solubility and, as L57 was expected to have poor water solubility (Table 1), it was discounted as an option for this study. Additionally, L95 peptide was considered unsuitable as it exhibited a relatively high positive charge (Table 1) and non-specific binding in vitro (Fig 2C); both of which are incompatible with selective delivery of the potent cytotoxic payload to be attached to this peptide. Furthermore, Auristatin was to be conjugated to a γ-aminobutanoate-MPA linker on the N-terminal of the peptide via a thiol-maleimide reaction. The linkers are expected to allow the peptide to bind to the receptor unhampered by the attached Auristatin. Therefore, a cysteine extra to the N-terminal cysteine, such as in L101peptide, means that a non-ideal mixed product of mono-conjugated and bi-conjugated peptide-Auristatin, may result. Receptor binding by a bi-conjugated product may be sterically hindered by the extra Auristatin. Considering all of these factors, and based on its predicted charge, solubility (Table 1), in vitro selectivity (Fig 2B and 2C) and similar in vivo performance to other assessed peptides (Fig 3B), we selected the LHa peptide as the preferred candidate for delivery of Auristatin to the testes.

Conjugation of menadione to FSH peptide and Auristatin to LHa peptide

In adapting these sequences for the purpose of carrying a specific payload, we attached a linker, comprised of γ-aminobutanoic acid and mercaptopropionic acid (Linker 2), that provided a reactive thiol group for attaching the peptide to other moieties or to nanoparticles via a maleimide-thiol coupling reaction [38, 39]. Additionally, the C-terminal of all peptides was amidated with the aim of inhibiting protease degradation, increasing stability [40, 41], and thereby prolonging the in vivo half-life and extending the time that the peptide is exposed to the target receptor. Additionally, we have previously shown that Linker 2, allows menadione to produce ROS by redox cycling despite its covalent attachment to the peptide [22]. The peptides with linker 2 are hereinafter referred to as LH2 and FSH2. FSH2 was attached, via the N-terminal cysteine, to C3 of menadione and the product, FSH2Menadione, isolated. FSH2Menadione had previously been synthesized, purified and the powdered product stored at -20°C [22]. LH2 was reacted with maleimide-Auristatin overnight and the product successfully purified by HPLC, freeze dried, and the powdered product stored at -20°C. The identity and purity of the product was confirmed by LC-MS that confirmed a product of mass ~2648 (representing LHa MW 1,333 + vcMMAE MW 1316). The reagents were resuspended in 30% Kolliphor/PBS immediately before administration to the treatment groups.

LH2Auristatin homed to the testes and impacted seminiferous tubule morphology and germ cell viability

Treated mice maintained healthy weight gain over a ten-week post-injection surveillance period, with no difference in mean body weights detected between the treatment groups. Additionally, with the exception of the testes (Fig 4A), there was no difference in tissue (seminal vesicles, epididymides, adrenals, kidneys, liver, brain, heart, and spleen) weights or morphology between the treatment groups, indicating that the injected peptides did not elicit pronounced off-target effects. In contrast, it appeared that the testes of at least some of the treated mice had been adversely impacted. Specifically, thirty percent of LH2Auristatin-treated males and an additional twenty percent of FSH2Menadione/LH2Auristatin-treated males presented with at least one testis that was significantly smaller (weights ranging from 41 up to 85 mg) than that of the controls. By comparison, all males that had been treated with either FSH2Menadione alone or the vehicle had testes weights over 126 mg. However, the overall mean weights of the testes of LH2Auristatin and FSH2Menadione/LH2Auristatin treated males were no different to that of the vehicle control group (Fig 4A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Effect of FSH2Menadione (FSH2Md) and LH2Auristatin (LH2Aur) in vivo.

Male adult mice were injected IP with 300 μl/30 g of either 14.5 mM FSH2Menadione; 100 μl/30 g 420 μM LH2Auristatin; a combination of both, 16 hours apart; or 300 μl/30 g of the vehicle (30% Kolliphor/PBS) with 10 mice in each treatment group. Five males in each treatment group were mated, with two control females each, six weeks post-injection for three weeks. Males were euthanized immediately after the mating period (~10 weeks post-injection) A. Average testis weights (n = 10). B. Tubule areas. H and E stained sections were imaged and then at least 50 tubules from each of five males from each treatment group was circled and the area in pixels/tubule was calculated, using imageJ (n = 5, p<0.05). Statistical analysis by One-way ANOVA then post hoc using student T-tests. C. Histology of the testes. Representative images of haematoxylin and eosin stained sections illustrating relative area of tubules. Scale bar = 50 μm.

https://doi.org/10.1371/journal.pone.0292198.g004

To account for the lower testes weights, the morphology of fixed testis sections was examined. Qualitatively, the seminiferous tubules of the LH2Auristatin- and the FSH2Menadione/LH2Auristatin-treated males appeared much smaller than those of the vehicle control or the FSH2Menadione-treated males (Fig 4B). Similarly, a quantitative evaluation using ImageJ software revealed that the average area occupied by seminiferous tubule in the testes of all of the treated animals were lower than that of the vehicle control group (Fig 4C). This reduction in tubule area proved highly significant (P < 0.005), such that the mice treated with LH2Auristatin and FSH2Menadione/LH2Auristatin exhibited a 40% reduction in mean tubule area compared to the testes of the control group. Such changes in tubule area were accompanied by the presence of apoptotic germ cells, which were clearly discerned in the two treatment groups (Fig 4B). To confirm germ cell apoptosis had indeed occurred, ApopTag was used to detect single-stranded and double-stranded breaks associated with this form of cell death. The results of this assay confirmed that the testes sections from both the LH2Auristatin and the FSH2Menadione/LH2Auristatin treatment groups exhibited significantly (P < 0.05) more tubules containing germ cells undergoing apoptosis (Fig 5A and 5B) and a significantly (P < 0.05) greater number of apoptotic cells per affected tubule (Fig 5C) than the FSH2Menadione alone or vehicle treatment groups Given the low number of apoptotic cells in the testes treated with FSH2Menadione alone, it appeared that the apoptosis was mostly precipitated by damage inflicted by the Auristatin.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Effect of FSH2Menadione (FSH2Md) and LH2Auristatin (LH2Aur) in vivo.

Male adult mice were injected IP with 300 μl/30 g of either 14.5 mM FSH2Menadione; 100 μl/30 g 420 μM LH2Auristatin; a combination of both, 16 hours apart; or 300 μl/30 g of the vehicle (30% Kolliphor/PBS) with 10 mice in each treatment group. Five males in each treatment group were mated, with two control females each, six weeks post-injection for three weeks. Males were euthanized immediately after the mating period (~10 weeks post-injection) A. Representative images of ApopTag. ApopTag staining was used to stain for cells undergoing apoptosis in sections from five males from each treatment group. Scale bar = 50 μm B. The total number of tubules/section (n = 5 sections/treatment group) and C. Percentage of tubules affected (>4 stained cells) were then counted along with the number of stained cells per affected tubule. NB. Subjects were randomly selected from the treatment groups but those testes weighing less than 60 mg were excluded. Statistical analysis by One-way ANOVA then post hoc using student T-tests.

https://doi.org/10.1371/journal.pone.0292198.g005

FSH2Menadione with LH2Auristatin had a small but not significant impact on Sertoli cell number

By targeting menadione to the testes using FSH2 peptide, we sought to ablate Sertoli cell function and vitality. To assess the success of this strategy, the number of Sertoli cells per seminiferous tubule were determined by immunolabeling of testes sections with antibodies directed against SOX9; a protein whose testicular expression is restricted to the Sertoli cell population. This analysis revealed a trend, albeit not statistically significant, of reduced mean numbers of Sertoli cells within the testes of the FSH2Menadione/LH2Auristatin treatment group (i.e. ~14 Sertoli cells/tubule) compared to the number of these cells in the vehicle control group (i.e. ~17 Sertoli cells/tubule) (S5B Fig). This trend was not replicated among Sertoli cell populations in the testes of males treated with either FSH2Menadione or LH2Auristatin, neither of which were numerically different from those treated with the vehicle. The lack of significant impact on the Sertoli cell population was reflected in the serum FSH levels, with an ELISA showing no difference in FSH levels within the cardiac sampled blood serum (S5C Fig) between the treatment groups. These data indicate that the targeted delivery of low dose menadione to Sertoli cells was not sufficient to create a significant reduction in their number.

The mature spermatozoa of FSH2Menadione- and FSH2Menadione/LH2Auristatin-treated males exhibited DNA strand breaks and oxidative DNA damage

Despite no overt effect on Sertoli cell populations, FSH2Menadione- and FSH2Menadione/LH2Auristatin-treatment did elicit oxidative DNA damage among the mature spermatozoa of treated males. Indeed, the application of a HALO assay indicated that there was a significant increase (P < 0.05) in DNA strand breaks harboured by the spermatozoa from FSH2Menadione- and FSH2Menadione/LH2Auristatin-treated animals (Fig 6A) compared to that of the control group. Specifically, an average of 22% of the spermatozoa from FSH2Menadione-treated and 17% from FSH2Menadione/LH2Auristatin-treated males displayed DNA strand breaks compared to only 5% of spermatozoa recovered from the cauda epididymis of LH2Auristatin- alone or vehicle-treated males. Additionally, there was a greater incidence of oxidation of DNA guanosine bases in the spermatozoa from the LH2Auristatin- and the FSH2Menadione/LH2Auristatin-treated treatment groups (averaged between 25 and 31%) compared to the vehicle-treated animals (Fig 6B). Furthermore, the oxidative DNA damage observed in the spermatozoa was evidently not a product of ongoing ROS production as there was no difference in the levels of either mitochondrial ROS or cytoplasmic superoxide detected within these cells (S6A and S6B Fig) between the four treatment groups. Moreover, the DNA damage burden in the caudal spermatozoa of treated males was not accompanied by a significant increase in the incidence of abnormal sperm morphology compared to the control males (Fig 6C). Computer-assisted sperm analysis (CASA) also showed no difference in sperm motility parameters or sperm total motility among the treated males compared to that of the vehicle controls (S7A–S7D Fig). Furthermore, there was no difference in the vitality of the caudal spermatozoa (S6C Fig) or in the DNA fragmentation index the sperm chromatin structure assay (SCSA) (S6D Fig) between the treatment groups. Notwithstanding these results, the DNA damage detected in the spermatozoa of some of the treatment groups appeared to be of sufficient magnitude to impinge on the fertility of the treated males, which was consequently assessed.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Epididymal sperm–effect of FSH2Menadione (FSH2Md) and LH2Auristatin (LH2Aur) in vivo.

Male adult mice were injected IP with 300 μl/30 g of either 14.5 mM FSH2Menadione; 100 μl/30 g 420 μM LH2Auristatin; a combination of both, 16 hours apart; or 300 μl/30 g of the vehicle (30% Kolliphor/PBS) with 10 mice in each treatment group. Males were euthanized 10 weeks post-injection and epididymal sperm was collected from all ten males in each treatment group and all samples were assessed for abnormal morphology, vitality, motility, mitochondrial ROS, cytoplasmic ROS, and DNA fractionation with at least one hundred sperm counted in each assay for each sample. A. Sperm were probed for DNA strand breaks using the HALO assay and sperm were categorised as either having a halo, or no halo. B. An anti-8-hydroxy-deoxyguanosine antibody was used to detect oxidised DNA guanosine bases. C. Sperm smears of the sperm were fixed in methanol and then stained using the Rapid Diff Kit (Pathtech). Sperm were categorised as either having normal or abnormal morphology. Statistical analysis by One-way ANOVA then post hoc using student T-tests.

https://doi.org/10.1371/journal.pone.0292198.g006

The combination of LH2Auristatin and FSH2Menadione decreased the fertility of treated sires

Eighty percent of the females that had been housed with the LH2Auristatin or the FSH2Menadione/LH2Auristatin injected males became pregnant with live embryos present at 13 days gestation. A mating plug was observed in the remaining 20% of females but no embryos were present at 13 days post-mating. All ovaries exhibited corpora lutea, indicating that ovulation was likely to have taken place in these females. There were significantly fewer average embryos/pregnant female mated with males from LH2Auristatin and FSH2Menadione/LH2Auristatin treatment groups compared to those females mated with the vehicle control sires (Fig 7A). Additionally, 87.5% of females mated with males in the FSH2Menadione/LH2Auristatin treatment group displayed evidence of embryo resorption, in the form of resorption moles, at 13 days post-mating (Fig 7B). Whilst all treatment groups, produced a higher than average number of resorption moles than that of the vehicle control group, the FSH2Menadione/LH2Auristatin treatment in particular produced a very high, 1.9 resorption moles/pregnant female (Fig 7B), with one female mated from this group exhibiting five resorption moles.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Effect of FSH2Menadione (FSH2Md) and LH2Auristatin (LH2Aur) in vivo on fertility.

Male adult mice were injected IP with 300 μl/30 g of either 14.5 mM FSH2Menadione; 100 μl/30 g 420 μM LH2Auristatin; a combination of both, 16 hours apart; or 300 μl/30 g of the vehicle (30% Kolliphor/PBS) with 10 mice in each treatment group. Five males in each treatment group were mated, with two control females each, six weeks post-injection for three weeks. At 13 days post-coitus, females were euthanized then embryos and resorption moles were counted. A. The number of embryos per pregnant female. B. The average number of resorption moles per pregnant female and the proportion of pregnant females that had a least one resorption mole.

https://doi.org/10.1371/journal.pone.0292198.g007

FSH2Menadione administration alone was not sufficient to compromise male fertility. In fact, all five males in the FSH2Menadione treatment group were fertile, with each of the ten females housed with these males becoming pregnant and carrying live embryos at 13 days post-mating (Fig 7A). Two of the females that had been mated with males from this treatment group exhibited one resorption mole each, with none detected in the remaining eight dams. In contrast to the above, all five males in the vehicle treatment group proved to be fertile with those females that became pregnant having at least 11 (up to 17) and an average of 13.5 embryos present at 13 days, none had resorption moles (Fig 7A).