TECHNICAL FIELD

The present invention relates to biomarkers of hypogonadism, in particular to male hypogonadism and thereto related endogenous molecular species being indicative of inadequate testosterone production.

BACKGROUND

Testosterone is the main androgen in males and its production is regulated by the hypothalamic- pituitary-gonadal axis. Currently, a low concentration of testosterone in blood is one major diagnostic factor for hypogonadism (testosterone deficiency syndrome) in male patients. However, the accurate and precise measurement of testosterone has remained a challenge as the testosterone activity has a weak correlation with its levels in plasma. Plasma testosterone levels are affected by diurnal, seasonal and aging variations. Furthermore, variations can result from medications, protein binding (e.g. Sex hormone binding globulin (SHBG) and albumin), BMI, and certain diseases (e.g. diabetes). Clinical signs and symptoms that indicate testosterone deficiency syndrome, though often unspecific, are currently at least as important as the available measurement technique of testosterone deficiency. Presently, the standard techniques to measure total testosterone are immunoassays and mass spectrometry (Taieb et al, Clin Chem., 2003, Aug 1, 49(8), 1381-95; and Cao et al, Clin Chim Acta, 2017, 469, 31-6), but these lead to diagnoses that are not very reliable and accurate for testosterone deficiency syndrome. It is well known that high doses of testosterone (and other anabolic steroids) are associated with adverse health effects. The association of testosterone with metabolic and cardiovascular function has gained increased attention, and it has been reported that testosterone deficiency, i.e. hypogonadism, is associated with increased mortality from cardiovascular disease and also from other causes (Vikan et al, European Journal of Endocrinology, 2009, 161(3), 435-42).

Hypogonadism afflicts almost 30% of men in the age range 46 to 89 years (Allan et al, Clinical Endocrinology, 2004, 60(6), 653-70). The condition is associated with changes in body composition (decreased lean body mass, osteoporosis, increased fat mass) and non-specific clinical symptoms are commonly displayed (Wu et al, New England Journal of Medicine, 2010, 363(2), 123-35). Testosterone deficiency also appears to be an independent risk factor for the development of metabolic syndrome (MetS) and type 2 diabetes (Araujo et al, Journal of Clinical Endocrinology and Metabolism, 2011, 96(10), 3007-19; and Vikan et al, European Journal of Endocrinology, 2009, 161(3), 435-42). The causality of the relationship between low testosterone and metabolic diseases is unclear, however, obesity-induced androgen deficiency and hypogonadism- induced obesity are both likely to bidirectionally contribute to disease pathology (Kelly et al., Journal of Endocrinology, 2013, 217(3), R25-45). There is a great unmet medical need of clinical markers, e.g. biomarkers, of testosterone action in order to improve the diagnosis of hypogonadism and monitoring common comorbidities such as, for example, cardiovascular disease, diabetes, metabolic syndrome, obesity, and prostate disease. SUMMARY

It is an objective of the present invention to diagnose hypogonadism with a higher level of certainty as compared to the corresponding diagnosis being based on measurement of systemic testosterone.

It is a related objective of the present invention to provide improved assessment of individual risk of one or several of cardiovascular disease, diabetes, metabolic syndrome, obesity and prostate disease.

These and other objectives, which will appear from the following description, have now been achieved by a method of diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof in an individual, the method comprising quantification of a plurality of biomarkers in a biological sample obtained from said individual, wherein said plurality of biomarkers are independently selected from the group of biomarkers consisting of ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol, according to an aspect of the present invention;

Further features of the invention and its embodiments are set forth in the appended claims.

It should be emphasized that the term “comprises/comprising” when used in this specification is taken to specify the presence of stated features, integers, steps or components but does not preclude the presence or addition of one or more other features, integers, steps, components or groups thereof.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects, features and advantages of which examples of the invention are capable of will be apparent and elucidated from the following description of examples of the present invention, reference being made to the accompanying drawings, briefly described below;

Fig. 1 (A) Study overview: mean value of the sex hormones of 30 volunteers across the three of sample collection time points. (B) 2D annotation enrichment analysis: biological process enriched for fold-change [B-A] versus fold-change [C-B]. (C) Heat map of the proteins with the highest degree of alteration between conditions. On top, the sex hormones are displayed as numerical annotations. On the right side: The tissue distribution. (D) Abundance range of proteins with the highest degree of alteration between conditions with available concentration values from the Plasma Proteome database (Nanjappa, Vishalakshi et al. 2014. “Plasma Proteome Database as a Resource for Proteomics Research: 2014 Update.” Nucleic Acids Research 42(D1): D959-65). Proteins that are components of some of the enriched biological processes are highlighted in yellow.

Fig. 2 General pipeline applied for discovering analytes as markers for testosterone activity. From Top-to-Down: Short-term controlled Human model: Thirty healthy young men (19 - 32 years) were included in the study. The subjects underwent a treatment plan in which, after the first blood collection (condition A), they received a subcutaneous injection with 240 mg of Gonadotropin Releasing Hormone antagonist (GnRHa). Three weeks later, after a second blood collection (condition B), they received an intramuscular injection with 1,000 mg of testosterone-undecanoate (Nebido®, Bayer Pharmaceuticals) and the blood samples were collected after two weeks (condition C). Biobank processing: Samples were automatically aliquoted and stored at -80 °C in a biobank at Lund University, Sweden. Analytical analysis: Ninety plasma samples from the 30 individuals in the three different time points were analysed separately by mass spectrometry, amino acid and routine clinical chemistry analysis. Statistical analysis: After selecting the analytes influenced by testosterone, ROC testing were performing to select the 15 top ranked analytes and its combination, capable for distinguishing testosterone values lower than 8 ng/mL. Fig. 3 Box-plots distribution and correlation vs testosterone for exemplary analytes influenced by testosterone, across conditions A, B and C. A: Upper, ten analytes influenced positively by testosterone (F9. CFD. IGFBP2, MINPP1, PI16, IGFBP6, Glu, Gly, Estradiol and PSA). Down, fifteen analytes influenced negatively by testosterone (FETUB, HPPD, CNDP1, ALDOB, Val, Met, Leu, Tyr, Phe, Lys, His, Trp, Calcium, ALAT, Urea). All represented analytes had significant changes between conditions A-B and B-C (FDR < 0.05), except for the amino acids Glu and Gly that only changed between conditions (B-C). B: Correlation of analytes vs testosterone across all different time points. Fig. 4 ROC testing results for all analytes influenced by testosterone. A: ROC-AUC distribution values of the 67 analytes testing for the separation of the group with low testosterone values (lower than 8ng/mL). The AUC is represented in percentage by bars, while the ROC significance thought of the p values is represented in an orange to red color scale. The analytes are also categorized by colors depending of the analytical technique used. The 15 top-ranked analytes based on the ROC-AUC values are represented on the top, with ROC-AUC values higher than 75%. The maximum sensitivity and specificity achieved for each analytes (top 15) are showed in brackets. B: ROC curves for four different combinations of markers. B-a: mass spectrometry (combined 1), B-b: amino acid analytes (combined 2), B-c: clinical chemistry analysis (combined 3). B-d: All analytes combined (combined 4).

DETAILED DESCRIPTION Embodiments of the present invention will be described in more detail below in order for those skilled in the art to be able to carry out the invention. The invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. The embodiments do not limit the invention, but the invention is only limited by the appended patent claims. Furthermore, the terminology used in the detailed description of the particular embodiments is not intended to be limiting of the invention.

Partial Studies

Below follow the disclosures of two partial studies: a first partial study and a second partial study. The below detailed first partial study and the second partial study disclose and teaches the setup and use of a testosterone deficient human model. By use of this model, the influence of testosterone on the protein profile, amino acids, and routine clinical chemistry biomarkers was determined. It was surprisingly found that a panel of analytes, or parts of such panel or individual analytes thereof, may be used as biomarkers of testosterone activity and/or related hypogonadism. Such a panel may comprise several of six proteins, eight amino acids and one hormone, as further detailed herein below.

The first partial study

To ascertain the influence of alterations in testosterone levels on the proteins profile of healthy individuals, a short-term controlled healthy human cohort was recruited. Thirty young men between the ages of 19 to 32 years and in good health were included in the study. Briefly, the subjects underwent a treatment plan whereby, after the first blood collection, a subcutaneous injection of 240 mg GnRHa (Degeralix®, Ferring Pharmaceuticals) was given in the abdominal region. Three weeks post-injection, after a second blood collection, an intramuscular injection of 1,000 mg of testosterone undecanoate (Nebido®, Bayer Pharmaceuticals) was given. Two weeks later, the last blood sample was collected. Hence, blood samples were collected from the 30 subjects at baseline (A), low testosterone (B), and restored testosterone (C) states. Testosterone, Follicle- stimulanting hormone (FSH) and Luteinizing hormone (LH) concentration was measured for all individuals at the three time points (see Figure 1A).

As a proof-of-concept, the 30 samples from each time point were combined into three large pools (A, B, and C). The three pools were processed, and analysed by liquid chromatography mass spectrometry (LCMS) using data-dependent acquisition (DDA).

The induced changes in testosterone and gonadrotopins provoked effects at the protein level. For each protein, the mean difference of expression between the conditions A-B and B-C was calculated. With the two new lists of mean differences we performed in Perseu a 2D analysis (GO and KEGG pathway annotation enrichment) which is based on a MANOVA (Cox, Juergen, and

Matthias Mann. 2012. “ID and 2D Annotation Enrichment: A Statistical Method Integrating Quantitative Proteomics with Complementary High-Throughput Data.” BMC bioinformatics 13 Suppl l(Suppl 16): S12.; Tyanova, Stefka, and Juergen Cox. 2018. “Perseus: A Bioinformatics Platform for Integrative Analysis of Proteomics Data in Cancer Research.” In Humana Press, New York, NY, 133-48) test. This analysis allowed the selection of proteins that are part of the same gene set (for example, associated with a biological process) and whose expression is overall higher or lower in a particular condition as compared with the others. Eightysix functional annotations were significantly-enriched (p.v < 0.05) (see Table S2A below).

[Table follows on next page] Table S2A. 2D enrichment analysis. Significantly-enriched functional annotations and pathways. Table S2A. 2D enrichment analysis. Significantly-enriched functional annotations and pathways. Table S2A. 2D enrichment analysis. Significantly-enriched functional annotations and pathways. Table S2A. 2D enrichment analysis. Significantly-enriched functional annotations and pathways.

[Text follows on next page] The 2D enrichment analysis (Figure IB) enabled division of the predominant enriched biological processes into four quadrants that were dependent on the protein changes between conditions. Quadrants II and IV represent the biological processes that are restored after administration of testosterone; whilst quadrant I and III show the processes that are only affected by gonadotropins. A clear enrichment of biological processes associated with disease was detected, e.g., diabetes type 2. This is related to the regulation of insulin-like growth factors, the innate inmune sytem, and the glucose metabolism (glycolysis/gluconeogenesis). Interestingly, IGF-I is known to be regulated by androgen (Keenan, B S et al. 1996. “Androgen Regulation of Growth Hormone Binding Protein.” Metabolism: clinical and experimental 45(12): 1521-26) and also, there is an association between the IGF-axis and testosterone-related diseases, e.g., osteoporosis, and cardiovascular disorders have been described (Kaplan, Robert C et al. “Insulin-like Growth Factors and Coronary Heart Disease.” Cardiology in review 13(1): 35-39.; Zofkova, I. 2003. “Pathophysiological and Clinical Importance of Insulin-like Growth Factor-I with Respect to Bone Metabolism.” Physiological research 52(6): 657-79). In addition, IGF-I enhances insulin sensitivity in healthy adults (Boulware, S D et al. 1994. “Comparison of the Metabolic Effects of Recombinant Human Insulin-like Growth Factor-I and Insulin. Dose-Response Relationships in Healthy Young and Middle-Aged Adults.” The Journal of clinical investigation 93(3): 1131-39; RUSSELL- JONES, D. L. et al. 1995. “A Comparison of the Effects of IGF-I and Insulin on Glucose Metabolism, Fat Metabolism and the Cardiovascular System in Normal Human Volunteers.” European Journal of Clinical Investigation 25(6): 403-11). Although insulin is the primary regulator of glucose metabolism, it has also been suggested that the IGF-axis plays a role in the glucose homeostasis process (Rajpathak, Swapnil N. et al. 2009. “The Role of Insulin-like Growth Factor-I and Its Binding Proteins in Glucose Homeostasis and Type 2 Diabetes.” Diabetes/Metabolism Research and Reviews 25(1): 3-12).

In our study, the glucose metabolism process (glycolysis/gluconeogenesis) was restored after the injection of testosterone, but in an inverse relationship to the IGF axis process (Figure 1C, IV). Other processes enriched in quadrant I, such as collagen degradation and angiogenesis, have also been associated with testosterone changes and related diseases (Chen, Yeping et al. 2012. “Testosterone Replacement Therapy Promotes Angiogenesis after Acute Myocardial Infarction by Enhancing Expression of Cytokines HIF-la, SDF-la and VEGF.” European journal of pharmacology 684(1-3): 116-24; Fukui, Michiaki et al. 2007. “Association between Serum Testosterone Concentration and Collagen Degradation Fragments in Men with Type 2 Diabetes Mellitus.” Metabolism: clinical and experimental 56(9): 1228-32; Yoshida, Sumiko, Yasumasa Ikeda, and Ken-ichi Aihara. 2016. “Roles of the Androgen— Androgen Receptor System in Vascular Angiogenesis.” Journal of atherosclerosis and thrombosis 23(3): 257-65). Based on the distribution of fold-changes between conditions, two probability curves were plot and 66 proteins with fold-change less than 5% probability (distribution tails) were selected (see Table S3A below).

[Table follows on next page]

Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails) Table S3 A. Proteins with fold-change less than 5% probability (distribution tails)

[Text follows on next page]

These proteins exhibited the highest degree of alteration between conditions and were selected as the most responsible for the observed functional changes. These were subsequently related to changes in testosterone and gonadotropins (see herein below). For these proteins, the entire abundance range is shown in Figure ID. Proteins with available concentration values from the Plasma Proteome database (Nanjappa, Vishalakshi et al. 2014. “Plasma Proteome Database as a Resource for Proteomics Research: 2014 Update.” Nucleic Acids Research 42(D1): D959-65) are also included. About four orders of magnitude in the dynamic range are represented (from 1.4 to 4800 ng/mL). The UniProtKB functional annotations (Cox, Juergen, and Matthias Mann. 2012. “ID and 2D Annotation Enrichment: A Statistical Method Integrating Quantitative Proteomics with Complementary High-Throughput Data.” BMC bioinformatics 13 Suppl l(Suppl 16): SI 2) are represented for some proteins.

Based on the expression pattern of the 66 proteins across the three time points studied, the heat map (Figure 1C) shows that B and C form a cluster likely because patients from these conditions have similar levels of gonadotropins. The tissue distribution of the proteins with the highest levels of altered expression displayed the heterogeneity of the proteins influenced by testosterone. Although more than 95% of these proteins (63 from 66) have been previously identified in plasma; only half (30 from 66) are predicted as secreted proteins in the Human Protein Atlas (https://www.proteinatlas.org/humanproteome/tissue/secretome ) (see Table S3A above). This results suggests that the remainder could be leakage proteins from tissues.

The results obtained indicated that it was indeed possible to reveal significant differences at the protein level across the three groups. Although this short-term human model is a controlled study with healthy subjects, the approach is considered a proof-of-concept for a much larger, highly-promising clinical evaluation. This is of particular importance because the causality of the relationship between low testosterone and metabolic diseases is not well-known. The approach was successful, but could benefit further from a deeper level of scrutiny. This would include the analysis of individual samples to take the such differences into account. In addition, future work on this approach will enable the identification of proteins that have a clinical and biological significance and could form the basis for the introduction of new biomarkers in routine healthcare. In conclusion, under the influence of altered levels of gonadotropin and testosterone in blood, the three conditions studied here were readily separated according to protein content and expression profile. Sample preparation and data analysis of the entitled “first partial study” above is described in detail below:

Thirty healthy young men between 19 - 32 years (mean = 23.9 years) with a BMI of 19.1 - 26.9 (mean = 23.0) were included in the study. Exclusion criteria included regular medication, exposure to steroids, drug use within the last year, stroke, heart and liver disease, cancer and any other chronic disease. The subjects visited Centre of Reproductive Medicine (RMC) at Skane University Hospital in Malmo, Sweden three times for the blood collection. The subjects underwent a treatment plan in which, after the first blood collection, they received a subcutaneous injection with 240 mg of Gonadotropin Releasing Hormone atagonist (GnRHa) (Degeralix®, Ferring Pharmaceuticals) in the abdominal region. Three weeks later, after a second blood collection, they received an intramuscular injection with 1,000 mg of testosterone-undecanoate (Nebido®, Bayer Pharmaceuticals). Two weeks later, the last blood collection took place. Hence, we collected blood samples from the 30 subjects at normal testosterone (A), low testosterone (B), and restored testosterone (C) states (see Figure 1). We measured Testosterone, FSH and FH for all subjets in the three time points. Samples were aliquoted and stored at -80 °C in a biobank at Fund University, Sweden (The study was approved by the Regional Ethical Review Board in Fund (http://epn.se/en), (Approval number: DNR 2014/311) and the participants signed informed consent. Sample order was randomized before starting sample preparation. Quantification of total proteins concentration in the samples was carried out by the Bicinchoninic Acid (BCA) assay methodl. The three pooled samples (30 samples in each one) were individually depleted with a MARS7 (Agilent, Santa Clara, CA) column following the instruction of the manufacturer (we took around 10 pF per sample). The buffer was exchanged to SDC 1.6%, 50 mM of Ambic using Amicon Ultra Centrifugal filter (0.5mF - 10 kDa, Millipore, Tullagreen, Ireland). The disulfide bonds were reduced by adding DTT to a final concentration of 10 mM and incubated 1 h at 37 °C. The free thiol groups were alkylated by adding iodoacetamide to a final concentration of 25 mM, and the reaction proceeded for an additional 30 min at room temperature in darkness (reduction and alkylation were done in the Amicon filter). The buffer was exchanged to Ambic 50 mM and the samples were resuspended in 100 pF of 50mM AmBic (30 pg of proteins after BCA quantification) and digested with trypsin at an enzyme-to-substrate mass ratio of 1/30, for 16 h at 37 °C. The remaining SDC was precipitated by adding 20% of formic acid prior to filtering the samples through a polypropylene filter plate with hydrophilic PVDF membrane (mean pore size 0.45 pm, Porvair Filtration Group). Samples for DDA were analyzed using a Q-Exactive Plus mass spectrometer connected to an Easy-nLC 1000 pump (Thermo Scientific, San Jose, CA) with a top 10 DDA or a PRM method (2 pL, 1 pg on the column). Peptides were loaded onto an Acclaim PepMap 100 precolumn (75 pm x 2 cm, Thermo Scientific, San Jose, CA), and separated on an easy-Spray column (25 cm x 75 pm ID, PepMap Cl 82 pm, 100 A) with the flow rate set to 300 nL/min and the column temperature to 35 °C.90 min gradients with Full MS scans were acquired with the Orbitrap mass analyzer over m/z 400-1600 and three more gas phase fractionations ranges (400-600, 590-900, 890-1600) with a resolution of 70,000 (at m/z 200). Target Automated Gain Control (AGC) value was set to le6 and maximum injection time of 100 ms. The ten most intense peaks with charge state > 2 were fragmented in the Higher-energy Collisional Dissociation (HCD) collision cell with a normalized collision energy of 26%. Tandem mass spectra were acquired in the Orbitrap mass analyser with a resolution of 35,000 (at m/z 200), target AGC value of 5e4 and maximum injection time of 100 ms. The underfill ratio was set to 10% and dynamic exclusion was 45 s.

We used Proteome Discoverer v 2.2 (Thermo Scientific, San Jose, CA) for peptide and protein identification. Peptides were identified using SEQUEST HT against UniProt human database (http://www.uniprot.org) integrated into Proteome Discoverer (Human 9606, Reviewed, 20 165). The search was performed with the following parameters applied: cysteine carbamidomethylation as a static modification, oxidation of methionine as a dynamic modification, 10 ppm precursor tolerance and 0.02 Da fragment tolerance. Up to one missed cleavage for tryptic peptides was allowed. The filters we used were high and medium confidence at peptide and protein level respectively, according to the Proteome Discoverer software. Up to one missed cleavage site for tryptic peptides was allowed. According to the Proteome Discoverer software, the filters applied were high- (FDR<0.01) and medium confidence (FDR<0.05) at peptide- and protein levels, respectively. The peptide/protein quantification was based on the quantification of MS peptide signals (Label-free quantification). Label-free quantification used the Minora Feature Detector node in the processing workflow, and the Precursor Ions Quantifier node and the Feature Mapper in the consensus workflow. Bioinformatics and statistical analysis were performed in Perseu2 and R 3,4softwares. Intensities values were normalized by log2 transformation and standardized by subtracting from each protein the median of the pooled sample. The corresponding gene ontology (GO) annotations (biological process, molecular function, and cellular compartment) and KEGG pathways were mapped to each protein using Perseu. For functional enrichment analysis, the 2D annotation enrichment algorithm created by Cox and Mann5 was applied. The lists of log2-fold-change between conditions A-B and B-C were used as input for the analysis and annotations with p. values <0.05 were considered significantly enriched.

The proteins with the highest degree of alteration between conditions were determined based on the log2-intensity differences (i.e. intensity ratios) between pools. Two probability distribution plots with the log2-intensity differences were made per comparison (A-B and B-C) and differences with less than 5 % of probability (higher ratios) were selected as significant in each case. We removed extreme values until obtain two normal distribution curves (Anderson-Darling testing). The cut-off defined to consider that a protein has a high degree of alteration between conditions A and B was ± 0.6 whilst between B and C was ± 0.5.

To visualize the changes of the proteins with the highest degree of alteration, a heat map based on protein expression together with a hierarchical clustering (distance: ‘correlation’; linkage: ‘complete’) was made. The concentration values of the sex hormones were used as numerical annotation to visualize their association with the protein expression clustering.

The Human Proteome Map (http://www.humanproteomemap.org/) was used to determine the tissue distribution of the DEPs proteins. The second partial study

In a previous pilot proteomics study using pooled plasma samples, we demonstrated that changes in the proteome profiles occur (the first partial study). In the current second partial study, 90 plasma samples were analyzed separately by mass spectrometry, amino acid analysis, and clinical routine chemical analytic assays. A set of analytes influenced by testosterone fluctuations is presented, of which we selected a panel of candidate markers for the diagnosis of testosterone activity.

Thirty healthy young men between the ages of 19 to 32 years were included in the study as previously taught in the first partial study. Unless otherwise stated, all chemical reagents were purchased from Sigma Aldrich (St. Louis, MO, USA). Modified porcine trypsin was purchased from Promega (Madison, WI, USA) and water was obtained from a Milli-Q ultrapure water system (Millipore, Billerica, MA, USA). Water and organic solvents for Liquid Chromatography Mass Spectrometry (LC-MS) were of LC-MS quality and supplied by Merck (Darmstadt, Germany).

The sample order was randomized prior to sample preparation. Quantitation of total protein was performed using the bicinchoninic acid (BCA) assay (Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH, Provenzano MD, et al. Measurement of protein using bicinchoninic acid. Anal Biochem. 1985;150(l):76-85). The plasma samples were processed and analyzed by liquid chromatography mass spectrometry (LC/MS). Sample preparation and the data analysis performed are described in detail in Table SIB below:

[Table follows on next page]

Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry

SUBSTI UTE SHEET (Rule 26) Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry Table SIB. Total proteins quantified by Mass Spectrometry

The plasma sample was mixed with an internal standard and the proteins precipitated by the addition of sulfosalicylic acid, pH was adjusted using lithium hydroxide. The separation of free amino acids in the supernatant was performed using a Biochrom 30+ amino acid analyzer system (Biochrom Ltd., Cambridge, UK). The compounds formed between ninhydrin and individual amino acids were quantified spectrophotometrically at 570 and 440 nm. Each amino acid was identified based on its retention time and the concentration automatically calculated using the software OpenLAB EZChrom. (see Table S2B below). Analytes from mass spectrometry techniques were pre-processed in Perseus vl.6.7.0 (Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, et al. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat Methods [Internet]. 2016 Sep 27 [cited 2018 Nov 7] ;13(9):731—40) software. The statistical analyses were performed using R software (Team R Core. R: A language and environment for statistical computing. [Internet]. Vienna, Austria: R Foundation for Statistical Computing.; 2018; RStudio Team. RStudio: Integrated Development for R. RStudio. [Internet]. Boston; 2016) and SPSS Statistics 21.0 (IBM, Somers, IL, USA). Analytes that changed significantly between conditions were determined by a one-way ANOVA considering the repeated nature of the samples (R function: ezANOVA{ez}). The ANOVA was followed by a paired Student's t-test (two-tails) as post hoc test, and adjusted (FDR) p-values < 0.05 were considered significant. Correlation between analytes and testosterone was determined considering the repeated nature of the data (R function: rmcorr{rmcorr}( Bakdash JZ, Marusich LR. Repeated measures correlation. Front Psychol. 2017 Apr 7;8(MAR))). Coefficients of correlation with p-value < 0.05 were considered significant. Since between conditions B and C the gonadotropins (FSH, LH) did not change but the testosterone did, analytes with significant changes between these two conditions were selected for later analyses as considered mainly influenced by testosterone changes. Receiver operating characteristic (ROC) curves were used as a discriminator system for the detection of low testosterone values using selected analytes. Two groups were defined: The low testosterone group (testosterone < 8 ng/mL) and the remaining values were considered normal. The ROC-AUC and binomial logistic regression for the combination of the analytes were performed using SPSS 25 and plotted using R software (ggroc{pROC|).

To the best of our knowledge, this is the first time that a panel for testosterone activity is reported. To achieve this, we designed a novel analytical pipeline based on a short-term controlled human model and high-throughput analysis of 90 plasma samples from 30 individuals across three different time points. An automatic sample processing, collecting, and storing of blood samples in a biobank, proteomics based on MS, amino acid and routine chemical analysis were applied. We analyzed a total of 746 analytes, of which 67 were significantly influenced by testosterone. Individuals receiver operating characteristic curves (ROC) were created in order to select the 15 best markers of testosterone changes (see Fig. 2).

Sixty-seven out of 746 analytes changed due to testosterone level modulation

(see Table S3B below): [Table follows on next page] Table S3B Table S3B Table S3B We separated the analytes in two main groups based on how they were affected by testosterone. The analytes that were positively influenced (n=41): increased or decreased in the same direction as testosterone, and the negatively influenced analytes (n=26): changed in opposite direction compared to testosterone. We selected for plotting some exemplary analytes from various analytical techniques, most of which had higher significant changes between conditions (Fig. 3A). The remaining box plots and all statistical information can be viewed in Table SIB and Table S3B. In addition, the correlation of the analytes vs testosterone across the three time points showed positives values for the analytes that decreased in condition B and their values were, in general, restored in condition C (influenced positively by testosterone), opposed to those values for which the expression increased in condition B, before restoring in C (Fig. 3B). Forty-six of 676 proteins were significantly influenced by testosterone (Table SIB, S3B). We showed in Fig. 3 A, B, how the intensity values for some of the proteins identified are spread out across the three different time points. Thirty-eight proteins were positively influenced by testosterone, such as the F9, CFD, IGFBP2, MINPP1, PI16 and IGFBP6 among others, while the remaining ten proteins were negatively influenced, for example the proteins FETUB, HPPD, CNDP1 and ALDOB (see Figure 3).

Thirteen out of 32 amino acids were significantly regulated by testosterone changes. The distribution of their intensities in all the conditions studied is shown in Fig. 3. Most of them, were positively influenced by testosterone (Phe, Tyr, Lys, among others). The only exceptions were glutamine (Gin) and glycine (Gly).

Eight out of 38 clinical biomarkers measured were selected as influenced by testosterone, estradiol, PSA (positively influenced) and two liver biomarkers: ALAT, AS AT among others, as shown in Fig. 3.

ROC analysis was used to quantify how accurately the selected analytes can discriminate between low testosterone (lower than 8 ng/mL (all values in condition B)) and the remaining values across the three time points studied. Eighty percent of the analytes (54 out 67) discriminated significantly the low testosterone group from the high one, considering the significant p-values associated to the ROC curve (see Fig. 4A). In order to be more stringent we selected the 15 top-ranked analytes based on the ROC-AUC values. Six proteins (ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA), eight amino acids (Phe, Tyr, Val, His, Trp, Met, Lys) and estradiol yielded the best performances after the ROC (ROC-AUC higher than 75%) and were selected for further analysis (Fig. 4A). The sensitivity ranges were from 62 to 100 %, while the specificities were from 67 to 95%. The p-values obtained for the top-ranked selected analytes were lower than 0.0001 after the ROC testing (all values reported in Table S3B).

In order to outperform the results obtained with the individual markers, we assessed multi-marker combinations using binomial logistic regression. For practical reasons, we first combined the analytes measured by the same technique. Model 1 combined five proteins (ALDOB, HPPD, PI16, MINPP1, and IGFBP6) obtained by mass spectrometry, model 2 combined eight amino acids (Phe, Tyr, Lys, Val, Met, Trp, Leu, and His) and model 3 combined PSA protein together with estradiol (see Fig. 4B). All combinations achieved ROC- AUC values higher than 94%, while the proteins from mass spectrometry and the amino acids analysis showed ROC-AUC values higher than 96% (combined 1 and 2). However, combined model 3 (PSA + estradiol) did not improve the sensitivity and specificity (100 % and 75 % respectively), compared to the independent analysis of estradiol (also 100 % and 75 %). As expected, the fourth combination (all 15 analytes together) achieved a ROC-AUC of 99.2% with maximum sensitivity and sensibility of 100 % and 97 % respectively (Fig. 4B- d).

We reported for the first time, a panel of blood analytes associated to the activity of testosterone. We designed a pipeline for high-throughput analysis, that included a short-term controlled human model, utilizing proteomics, amino acid and routine clinical chemistry data. The analysis of the top-ranked 15 analytes could be useful as complementary clinical markers for diagnosing hypogonadism or testosterone deficiency.

In addition to our experimental design, only three studies have been published based on short-term effects of testosterone changes in healthy males (Rabiee A, Dwyer AA, Caronia LM, Hayes FJ, Yialamas MA, Andersen DK, et al. Impact of acute biochemical castration on insulin sensitivity in healthy adult men. Endocr Res. 2010;35(2):71-84; Host C, Gormsen LC, Hougaard DM, Christiansen JS, Pedersen SB, Gravholt CH. Acute and short-term chronic testosterone fluctuation effects on glucose homeostasis, insulin sensitivity, and adiponectin: A randomized, double-blind, placebo-controlled, crossover study. J Clin Endocrinol Metab. 2014;99(6): 1088-96), neither of them was focused on the large scale analysis of analytes for testosterone activity. The experimental designs were oriented to the comparison of individuals before and after castration in order to determine, mainly the effect on glucose metabolism and/or insulin sensitivity. Six proteins and nine metabolites (8 amino acids + estradiol) were selected as markers for testosterone activity. Notably, PSA was included (ROC-AUC= 76 %, Sensitivity =70% and Specificity=78%), although a controversial analysis has been reported relating PSA and testosterone concentrations (Corona G, Boddi V, Lotti F, Gacci M, Carini M, De Vita G, et al. The Relationship of testosterone to prostate-specific antigen in men with sexual dysfunction. J Sex Med. 2010;7(1 PART l):284-92). However, a previous study reported PSA as possible biomarker for testosterone deficiency with a sensitivity and specificity of 65.2% and 55.5%, respectively whereas almost 3000 patients without prostate disease and with PSA levels < 4 ng/mL were studied (Rastrelli G, Corona G, Vignozzi L, Maseroli E, Silverii A, Monami M, et al. Serum PSA as a predictor of testosterone deficiency. J Sex Med. 2013;10(10):2518-28). Our results supported this conclusion and also demonstrated that using our short-term controlled study it is possible to find and extrapolate significant changes in the reported analytes. Interestingly, another prostate related protein was included in our final set of analytes: the peptidase inhibitor 16 protein (MINPP1). Although, we did not find any previous relation of this protein with testosterone, it has been reported as a novel and independent prognostic markers following radical prostatectomy for prostate cancer (Reeves JR, Dulude H, Panchal C, Daigneault L, Ramnani DM. Prognostic value of prostate secretory protein of 94 amino acids and its binding protein after radical prostatectomy. Clin Cancer Res. 2006 Oct 15;12(20 PART l):6018-22).

As expected, estradiol was also selected since more than 80 % of circulating estradiol in men is derived from the aromatization of testosterone (Longcope C, Kato T, Horton R. Conversion of blood androgens to estrogens in normal adult men and women. J Clin Invest. 1969;48(12):2191-201). Both deficiencies (testosterone and estradiol) have also independent clinical manifestations for patients with hypogonadism(Finkelstein JS, Lee H, Bumett-Bowie S-AM,

Pallais JC, Yu EW, Borges LF, et al. Gonadal Steroids and Body Composition, Strength, and Sexual Function in Men A BS T R AC T. N Engl J Med. 2013 ;369: 1011-33; Rochira V, Carani C. Estrogen Deficiency in Men. In 2017. p. 797-828). Furthermore, estradiol has been reported as a better predictive marker than testosterone of osteoporosis in men over the age of 50 years (Clapauch R, Mattos TM, Silva P, Marinheiro LP, Buksman S, Schrank Y. Total estradiol, rather than testosterone levels, predicts osteoporosis in aging men. Arq Bras Endocrinol Metabol. 2009;53(8): 1020-5).

Three proteins from the Insulin growth factor axis (IGF-axis) were influenced positively by testosterone (IGFBP2, IGFBP5 and IGFBP6), although only

IGFBP6 was top-ranked and finally selected. IGF-axis is known to be regulated by androgen (Keenan BS, Richards GE, Mercado M, Dallas JS, Eakman GD, Baumann G. Androgen regulation of growth hormone binding protein. Metabolism [Internet]. 1996 Dec [cited 2019 Oct 9] ;45(12): 1521— 6) and to be associated with testosterone-related diseases, e.g., osteoporosis, and cardiovascular disorders (Zofkova I. Pathophysiological and clinical importance of insulin-like growth factor-I with respect to bone metabolism. Physiol Res [Internet]. 2003 [cited 2019 Oct 9];52(6):657-79; Kaplan RC, Strickler HD, Rohan TE, Muzumdar R, Brown DL. Insulin-like growth factors and coronary heart disease. Cardiol Rev [Internet], [cited 2019 Oct 9]; 13(1): 35-9). In addition, together with HPPD, the genes were found to be significantly regulated in a gene expression signature of human spermatogenic failure with distinct stages of male germ cell development using biopsies from men with highly defined and homogenous testicular pathologies (Spiess AN, Feig C, Schulze W, Chalmel F, Cappallo-Obermann H, Primig M, et al. Cross-platform gene expression signature of human spermatogenic failure reveals inflammatory-like response. Hum Reprod. 2007;22(ll):2936-46) (Ref.dataset: E-TABM-234). Multiple inositol polyphosphate phosphatase 1 (MINP1) protein regulates cellular levels of inositol pentakisphosphate (InsP5) and inositol hexakisphosphate (InsP6) and not much information about their relationship with testosterone or any other androgens was found. However, the MINP1 protein has been related with bone mineralization and ossification(Caffrey JJ, Hidaka K, Matsuda M, Hirata M, Shears SB. The human and rat forms of multiple inositol polyphosphate phosphatase: Functional homology with a histidine acid phosphatase up-regulated during endochondral ossification. FEBS Fett. 1999 Jan 8;442(1):99-104), which are processes related to osteoporosis, a bone related disease observed with decreased testosterone levels(Mohamad NV, Soelaiman IN, Chin KY. A concise review of testosterone and bone health. Clin Interv Aging. 2016;11:1317-24). Aldolase B is mainly produced in the liver and has been strongly associated with hepatic gluconeogenesis, because fructose is almost entirely metabolized in the liver in humans, where it is directed toward renewal of liver glycogen. In turn, the gluconeogenesis process has been strongly associated with hypogonadism and obesity (Aoki A, Fujitani K, Takagi K, Kimura T, Nagase H, Nakanishi T. Male Hypogonadism Causes Obesity Associated with Impairment of Hepatic Gluconeogenesis in Mice. Biol Pharm Bull [Internet]. 2016 [cited 2019 Oct 31] ;39(4):587— 92; Chevalier S, Burgess SC, Malloy CR, Gougeon R, Marliss

EB, Morais JA. The greater contribution of gluconeogenesis to glucose production in obesity is related to increased whole-body protein catabolism. Diabetes. 2006;55(3):675-81).

Most of the amino acids selected have been observed in previous studies of hypogonadism or testosterone related disease as diabetes type II (Fanelli G, Gevi F, Belardo A, Zolla L. Metabolic patterns in insulin-sensitive male hypogonadism, [cited 2019 Apr 24]; Wang TJ, Larson MG, Vasan RS, Cheng S, Rhee EP, McCabe E, et al. Metabolite profiles and the risk of developing diabetes. Nat Med. 2011 Apr;17(4):448-53). Notably, the two aromatic amino acids (Phe, Tyr) were the most influenced by testosterone in our study. Tyrosine was reported significantly up-regulated in hypogonadism males versus controls, while both amino acids were among the best predictors of risk of developing diabetes (Wang TJ, Larson MG, Vasan RS, Cheng S, Rhee EP, McCabe E, et al. Metabolite profiles and the risk of developing diabetes. Nat Med. 2011 Apr;17(4):448-53). Interestingly, the enzyme HPPD, i.e. 4- hydroxyphenylpyruvate dioxygenase, belonging to the Phenylalanine catabolism pathway was affected by testosterone fluctuations and was in turn selected as one of the top-ranked proteins.

Although each selected analyte could be independently able to discriminated testosterone deficiency, the combination of them could be used as complement for diagnosing hypogonadism. The first two combined models (Fig. 4B, combined 1 and 2) allowed ROC-AUC values higher than 94% with sensitivities and specificities higher than 90 % and 85% respectively. In the case of the combination between PSA and estradiol (combined 3), the specificity was under 85%. Consequently, combined model 3 should be considered unpractical at least with the data analyzed in the present study.

Because new factors or comorbidities could be added when patients are diagnosed, the analysis of all markers should be tested individually for specific cohorts. For amino acid analysis, the measurement of Phe or Tyr is enough in order to reach more than 90% of ROC-AUC value, however, all amino acids can be easily measured in the same experiment. The same concept applies for the proteomics based on mass spectrometry experiment, where all analytes can be analyzed in one multi-testing approach.

The limitations of our study are mainly associated to the sample size, because it is relatively small. In addition, the subjects included are healthy young men and our system is relatively ideal. Patients diagnosed with hypogonadism usually exhibit comorbidities, including e.g. cardiovascular disease, diabetes, metabolic syndrome, obesity, osteoporosis and prostate disease, which makes comparison with our model difficult. At the same time, the experimental design is one of the strengths of our study, since it allows us to differentiate the effect of testosterone with minimal interference. The values reported for each analyte are mostly based on relative quantifications because the main objective of this report is to demonstrate how well low testosterone values could be discriminated. These analytes should be tested in disease cohorts to determine the best cut-off in large number of samples. In conclusion, we reported a set of analytes influenced by testosterone fluctuations developed for high-throughput analysis of plasma samples. This set should be tested as complement for diagnosing hypogonadism or testosterone deficiency in patients with other diseases.

Decription The present invention relates to a method of diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof in an individual, such as e.g. a patient. Such comorbidities may include, for example, cardiovascular disease, diabetes, metabolic syndrome, obesity, and prostate disease. The qualitative or preferably quantitative (relative or absolute) determination of at least two biomarkers in a biological sample, such as e.g. a blood- or plasma sample, of an individual, such as a male individual, is an essential step of the present invention. The biomarkers may be proteins, such as e.g. one or several of ALDOB, HPPD, IGFBP6, PI16, MINPP1 and PSA, one or several of total individual amino acids of the proteins of the sample such as e.g. Phe, Tyr, Val, His, Trp, Met, Lys, or the hormone estradiol. Analytical methods for the detection and quantitative determination of such biomarkers are well known in the art. Proteins may for example be determined by mass spectrometry, in their native, modified or cleaved forms. The amino acids of the sample may be determined, for example, by initial hydrolysis of the corresponding proteins followed by quantification by a suitable automatic amino acid analyzer.

The pharmacological effect of testosterone is exercised through the interaction with target proteins that are identical between men and women. Hence, the following down-stream pathways overlap in men and women, although they may be attenuated differently due to other sex related factors. The absolute threshold levels of biomarkers indicative of male and female hypogonadism may differ between men and women, but their qualitative nature is preserved between sexes. Special conditions exist in prepubertal children who have a very low but still measurable production of sex hormones. The same reasoning may be conducted for different sub-groups within the group of mammals.

According to one embodiment, two of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol.

According to one embodiment, three of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol.

According to one embodiment, four of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol.

According to one embodiment, five of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol.

According to one embodiment, six of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol.

According to one embodiment, all of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met, Lys and estradiol.

According to one embodiment, all of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, IGFBP6, PI 16, MINPP1, PSA, Phe, Tyr, Val, His, Trp, Met and Lys. According to one embodiment, all of the following biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB, HPPD, Phe and Tyr. According to one embodiment, the following two biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of hypogonadism or prognosing a comorbidity thereof: ALDOB and HPPD.

According to one embodiment, the following two biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of male hypogonadism or prognosing a comorbidity thereof: ALDOB and HPPD.

According to one embodiment, the following two biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of female hypogonadism or prognosing a comorbidity thereof: ALDOB and HPPD.

According to one embodiment, the following two biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of male hypogonadism or prognosing a comorbidity thereof in a mammal such as a human: ALDOB and HPPD. According to one embodiment, the following two biomarkers may be quantified for diagnosing or prognosing or monitoring or staging the progression of female hypogonadism or prognosing a comorbidity thereof in a mammal such as a human: ALDOB and HPPD.

According to one embodiment, the method of the invention may be used for the determination of relative probability or risk of an individual to get a future diagnosis of one are several of the following deceases or conditions: cardiovascular disease, diabetes, metabolic syndrome, obesity and prostate disease.

In the claims, the term “comprises/comprising” does not exclude the presence of other elements or steps. Furthermore, although individually listed, a plurality of means, elements or method steps may be implemented. Additionally, although individual features may be included in different claims, these may possibly advantageously be combined, and the inclusion in different claims does not imply that a combination of features is not feasible and/or advantageous. In addition, singular references do not exclude a plurality. The terms “a”, “an”, “first”, “second” etc do not preclude a plurality.