Small lipidated anti-obesity compounds derived from neuromedin U

Abstract

A small library of truncated/lipid-conjugated neuromedin U (NmU) analogs was synthesized and tested in vitro using an intracellular calcium signaling assay. The selected, most active analogs were then tested in vivo, and showed potent anorexigenic effects in a diet-induced obese (DIO) mouse model. The most promising compound, NM4-C₁₆ was effective in a once-weekly-dose regimen. Collectively, our findings suggest that short, lipidated analogs of NmU are suitable leads for the development of novel anti-obesity therapeutics.

Graphical abstract

1. Introduction

Obesity is a leading preventable cause of death worldwide, with increasing prevalence in adults and children. This has become one of the most serious public health problems of the 21^st century [1]. Generally, obesity increases the likelihood of various diseases, particularly heart disease, type 2 diabetes, obstructive sleep apnea, certain types of cancer, and osteoarthritis [2]. Currently, therapeutics capable of reducing appetite and increasing energy expenditure are of high interest, and gut peptides which regulate energy homeostasis represent attractive leads [3–5]. Neuromedin U (NmU) is an endogenous peptide, highly conserved across species, that is implicated in a number of physiological processes including nociception, stress, inflammation, blood pressure, feeding, energy homeostasis, and glycemic control [6,7]. NmU is widely distributed in the body both peripherally and centrally [8–11], and its function is mediated by two G-protein coupled receptors (GPCRs): NMU1 and NMU2 [6,12–18]. NMU1 is predominantly expressed in the peripheral tissues, particularly the gastrointestinal tract, pancreas, uterus and testes [16,18], whereas NMU2 is mainly expressed in the central nervous system (CNS) with the highest levels in the hypothalamus, hippocampus, spinal cord and paraventricular nucleus [11,14,15]. In humans, NMU gene variants have been linked to excess body weight [19].

The bioactivity of NmU is mediated mainly through its C-terminal conserved region: F¹L²F³R⁴P⁵R⁶N⁷-amide [6,20]. To date, several structure–activity relationship (SAR) studies of this fragment have been carried out, establishing structural requirements for potent agonistic activity [21–28]. Generally, R⁶ and N⁷-amide are necessary for binding and activation respectively of the avian peripheral receptor and substitution of F³→Y results in improved bioactivity. N-terminal modifications with pyroglutamic acid, succinic acid and glutaric acid are permissive, showing improved aminopeptidase resistance and increased agonistic activity in contractility assays [26]. Using the same assays it was also determined that substitution of either, F¹, F³, R⁴, P⁵, R⁶, or N⁷-amide with glycine or their D-amino acid counterparts results in decreased bioactivity [29]. At murine NMU receptors, residues R⁴ and R⁶ are necessary for the agonistic activity with R⁴ being critical [30]. Moreover, the C-terminal N⁷-amide moiety is required for intracellular Ca²⁺ signaling by both recombinant human NMU1 [13] and murine NMU1 and NMU2 [30]. However recently published data for modified human full-length and truncated NmU analogs [31] showed that the N-terminal N⁷-amide may be replaced by D-norleucine-amide, giving additional selectivity toward NMU2 (compounds K and Q). Similar selectivity toward NMU2 is also observed for R⁶→homoArg replacement in combination with F¹→W or F¹→W and P⁵→A substitutions (compounds I, J, O and P). Notably, other receptor-selective analogs of NmU were also recently described [28]. The selectivity toward NMU2 was achieved by simultaneous modifications in positions F¹, F³ and R⁴. Similar selectivity toward NMU1 was exerted by modifications of L² at both its side chain and α–amine-group. The modification of P⁵ can also have variable outcome in terms of bioactivity. For example, in compound 1b [28] such modifications always lead to analogs with decreased activity regardless of substituent (e.g. Hyp, compound S4a; L-pipecolic acid, compound S4b; and 1-aminocyclopropane-1-carboxylic acid, compound S4c). On the other hand in certain modification schemes they seem beneficial (compounds J and P) [31].

Peripheral administration of NmU reduces food intake and body weight in rodents and birds [15,32–35], as well as increasing locomotor activity and core body temperature in rodents [34] with effects primarily being mediated by NMU2 [36]. Moreover, NmU-overexpression in mice results in a lean phenotype with improved glucose homeostasis [37]. In contrast, NmU-deficient mice develop obesity characterized by hyperphagia, reduced energy expenditure and hyperglycemia [38]. Notably, the chronic administration of NmU does not cause tachyphylaxis and significantly improves glucose tolerance in diet-induced obese (DIO) mice [39]. The unfavorable pharmacokinetic properties of NmU (the half-life of NmU after subcutaneous (s.c.) injection is less than 5 min [39]) were improved by conjugation with polyethylene glycol (PEG) [40] or human serum albumin (HSA) [41] showing in both cases long-lasting, potent anorectic, and blood glucose-normalizing activity. Although mentioned in a patent [31], full length lipid-conjugated (NMU24) and lipid/PEG40-bis-conjugated (NMU25) analogs of neuromedin U showed limited activity. Interestingly, centrally administered NmU-related neuropeptide neuromedin S (NmS), which shares the C-terminal sequence with NmU and activates the same receptors, exerts even greater anorexigenic effect through the corticotropin releasing hormone (CRH) and α-melanocyte-stimulating hormone systems [36,42].

2. Results

Lipidated and additionally stabilized short peptides derived from hormones have been shown to be suitable agents for hormone-replacement therapy [43–46]. To test whether NmU could be modified/derivatized to yield a similar type of compound(s), we synthesized a small library (Table 1) of truncated NmU analogs utilizing various stabilization protocols [47], including: cyclization, lipidation, introduction of α, α-disubstituted amino acids, a retro-inverso-approach, and a combination of these. All peptides were synthesized as C-terminal amides using a standard Fmoc protocol [48] and characterized by analytical RP-HPLC and MALDI-MS (Table 1 & Figure S1). Generally lipidation was achieved by N-terminal conjugation of palmitic acid. However in the case of retro-inverso derivatives and analog NM4A we used iminodiacetic acid mono-N-palmityl amide (Ida^NHPal, see Figure 1), which we previously found to be a useful lipidation moiety [44,45]. The double-lipidated analogs, NM4A-C₁₆ and NM4-C₁₆, were synthesized by reductive alkylation using a previously described protocol [49]. As starting materials NM4A and Ahx-Aib-FLFRPRN-amide were employed respectively. Reaction(s) were carried out “in solution” (1,4-dioxane:CH₃OH:H₂O/5:4:1) with an excess of 1-hexadecanal (50 eq) and sodium cyanoborohydride (NaBH₃CN, 100 eq) as the reductive agent. The cyclic analogs NM11-NM16 were synthesized from their linear counterparts (NM8-NM10) using a published S-alkylation protocol [50]. Reaction(s) were carried out in a 50% solution of DMSO in DMF in the presence of cesium carbonate (Cs₂CO₃) and tetrabutylammonium iodide (TBAI), producing all of the expected analogs but with disappointingly low yields (<5%).

Table 1.

Sequences, analytical and in vitro activity data obtained for NM analogs.

Peptide	Sequence	Composition	MW Calc / Found	RT (min)	NMU1			NMU2
					pEC50	Emax (% hNmU)	Emax at 10 μM as % hNmU Emax	pEC50	Emax (% hNmU)	Emax at 10 μM as % hNmU Emax
hNmU	FRVDEEFQSPFASQSRGY -FL- Phe -Arg- Pro-R-Asn	C141H203N41O38	3080.41 / 3080.91	32.23	8.97±0.12	100		9.10±0.16	100
NM1	Ac -FL- Phe -Arg- Pro-R-Asn	C47H71N16O9	990.17 / 990.96	30.74	9.24±0.12	45.5±4.3	-	9.47±0.17	135.2±0.2	-
NM2	Aib -FL- Phe -Arg- Pro-R-Asn	C49H76N16O9	1033.24 / 1034.14	25.77	9.25±0.22	66.3±0.6	-	9.67±0.19	69.9±0.1	-
NM3	Pal-Aib -FL- Phe -Arg- Pro-R-Asn	C65H106N16O10	1271.65 / 1272.79	64.44	6.85±0.08	87.8±1.4	-	6.97±0.30	126.2±0.2	-
NM31	Pal-Aib -FL- Phe-NMeR- Pro-R-NMeN	C67H110N16O10	1299.71 / 1300.86	57.24*	-	-	13.8±2.0	-	-	17.9±1.5
NM32	Pal-Aib -FL- NMeF -Arg- Pro-R-NMeN	C67H110N16O10	1299.71 / 1300.98	57.16*	-	-	19.1±1.2	-	-	15.2±3.8
NM33	Pal-Aib -FL- αMeF -Arg- Pro-R-NMeN	C67H110N16O10	1299.71 / 1300.89	56.22*	-	-	20.6±1.9	-	-	19.2±2.0
NM34	Pal-Aib -FL- Phe-NMeR- Pro-R-NMeN	C67H110N16O10	1299.71 / 1301.14	56.41*	-	-	19.6±3.2	-	-	18.4±3.2
NM35	Pal-Aib -FL- NMeF -Arg- Pro-R-NMeN	C67H110N16O10	1299.71 / 1300.96	55.85*	-	-	22.9±3.8	-	-	67.0±16.5
NM36	Pal-Aib -FL- αMeF -Arg- Pro-R-NMeN	C67H110N16O10	1299.71 / 1301.11	55.13*	-	-	15.9±6.6	-	-	28.6±7.9
NM3D1	CT-IdaNHPal -FL- Phe -Arg- Pro-R-Asn-NT	C65H107N17O10	1286.67 / 1287.44	50.30*	-	-	-	-	-	-
NM3D2	CT-IdaNHPal-Ahx -FL- Phe -Arg- Pro-R-Asn-NT	C71H118N18O11	1399.83 / 1400.21	49.76*	-	-	-	-	-	-
NM3D3	CT-IdaNHPal-Ahx-Aib -FL- Phe -Arg- Pro-R-Asn-NT	C75H125N19O12	1484.93 / 1485.29	50.34*	-	-	-	-	-	-
NM4	Pal-dPEG12-Aib -FL- Phe -Arg- Pro-R-Asn	C92H159N17O23	1871.37 / 1872.04	58.82	8.75±0.15	99.3±15.5	-	8.72±0.10	92.8±2.2	-
NM4A	IdaNHPal-Aib -FL- Phe -Arg- Pro-R-Asn	C69H114N18O11	1371.77 / 1372.21	49.28*	7.91±0.45	55.7±10.0	-	7.99±0.09	67.2±0.1	-
NM4A-C16	C16-IdaNHPal-Aib -FL- Phe -Arg- Pro-R-Asn	C85H146N18O11	1596.18 / 1596.36	63.19*	~ 4.9Δ	-	118.0±29.6 #	~ 5.0Δ	-	99.6±17.9 #
NM4A-Nic	Nic-IdaNHPal-Aib -FL- Phe -Arg- Pro-R-Asn	C75H117N19O12	1476.85 / 1476.89	53.64*	7.32±0.18	57.7±14.8	-	7.37±0.09	52.8±0.2	-
NM4-C16	(C16)2-Ahx-Aib -FL- Phe -Arg- Pro-R-Asn	C87H151N17O10	1595.24 / 1595.28	63.70*	~ 6.3Δ	-	81.0±3.6 #	~ 6.1Δ	-	28.4±4.3 #
NM5	Aib -FL- Phe -Arg- Pro-R-Asn-Aib	C53H83N17O10	1118.34 / 1118.62	26.22	7.23±0.09	77.0±4.6	-	7.25±0.09	104.7±0.2	-
NM6	Pal-Aib -FL- Phe -Arg- Pro-R-Asn-Aib	C69H113N1O11	1356.76 / 1356.82	64.48	-	-	40.6±2.2	-	-	60.7±12.1
NM7	Pal-dPEG12-Aib -FL- Phe -Arg- Pro-R-Asn-Aib	C96H166N18O24	1956.48 / 1956.82	58.79	6.89±0.12	77.8±4.7	-	6.97±0.26	66.5±0.1	-
NM8	C-Ahx -FL- Phe -Arg- Pro-R-Asn-Ahx-C	C63H101N19O12S2	1380.73 / 1381.99	30.65	6.79±0.50	59.8±0.5	-	6.55±0.10	53.6±0.1	-
NM9	C-Ahx -FL- Phe -Arg- Oic-R-Asn-Ahx-C	C67H107N19O12S2	1434.82 / 1435.46	33.00	5.87±0.08	50.1±1.3	-	6.14±0.13	35.5±0.3	-
NM10	C-Ahx -FL- Phe -Arg- Tic-R-Asn-Ahx-C	C68H103N19O12S2	1442.80 / 1443.57	33.69	~ 5.3Δ	-	61.0±12.7	~ 5.6Δ	-	97.8±11.8
NM11	X(C-Ahx -FL- Phe -Arg- Pro-R-Asn-Ahx-C)	C71H107N19O12S2	1482.87 / 1484.28	33.69	-	-	33.7±0.2	-	-	60.6±11.6
NM12	X(C-Ahx -FL- Phe -Arg- Oic-R-Asn-Ahx-C)	C75H113N19O12S2	1536.96 / 1538.37	35.33	-	-	30.00±6.0	-	-	70.0±6.0
NM13	X(C-Ahx -FL- Phe -Arg- Tic-R-Asn-Ahx-C)	C76H109N19O12S2	1544.94 / 1545.62	35.93	-	-	60.4±13.8	-	-	39.3±8.8
NM14	PalS-X(C-Ahx -FL- Phe -Arg- Pro-R-Asn-Ahx-C)	C88H141N19O12S3	1753.38 / 1753.97	60.53	5.87±0.05	50.0±3.0	-	6.00±0.09	53.1±0.3	-
NM15	PalS-X(C-Ahx -FL- Phe -Arg- Oic-R-Asn-Ahx-C)	C92H147N19O12S3	1807.47 / 1807.82	60.77	~ 5.2Δ	-	96.4±7.7	~ 5.2Δ	-	118.4±4.8
NM16	PalS-X(C-Ahx -FL- Phe -Arg- Tic-R-Asn-Ahx-C)	C93H143N19O12S3	1815.45 / 1816.37	61.51	~ 5.1Δ	-	133.8±22.0	~ 5.2Δ	-	126.6±10.8

All peptides were synthesized as C-terminal amides, NM3D1-D3 are all (D)-retro-inverso-analogs where: NT is the N-terminus and CT is the C-terminus of the peptide(s). Abbreviations: Ahx-6-aminohexanoic acid; Aib-aminoisobutyric acid; Ida-iminodiacetic acid, α^MeF-α-methyl-L-phenylalanine; N^MeR-N^α-methyl-L-arginine; N^MeN-N-methyl-L-asparagine; N^MeF-N-methyl-L-phenylalanine; Oic-(2S,3aS,7aS)-octahydro-1H-indole-2-carboxylic acid; Tic-(3S)-1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid; Pal-palmitic acid; PalS-X-3,5-bis(methyl-S-cysteinyl)-1-(methyl-S-palmityl)-benzene; X-1,3-bis(methyl-S-cysteinyl)-benzene; NHPal- N-palmitylamide; dPEG₁₂-40-amino-4,7,10,13,16,19,22,25,28,31,34,37-dodecaoxotetradecanoic acid. Analytical RP-HPLC profiles were obtained with an analytical reversed-phase C₁₈ SymmetryShield^™ RP18 column, 4.6×250 mm, 5 μm (Waters Corp., Milford, MA) or ⁽*⁾ an analytical reversed-phase XBridge^™ BEH300 C₄ column, 4.6×150 mm, 3.5 μm (Waters Corp., Milford, MA). pEC₅₀ and E_max values were obtained using a calcium signaling assay where data are mean ± SEM, n≥4.

^#E_max at 30 μM as % hNmU E_max.

^ΔApproximate EC₅₀ values were derived from curves that had insufficient data points at high concentrations of ligand but which did approach the E_max of hNmU. Values are presented to give an indication of likely agonist potency.

Figure 1.

The most active analogs of NmU synthesized for this study: (A) NM4, (B) NM4A and (C) NM4-C₁₆.

All synthesized NmU-analogs were tested in vitro using a cellular calcium signaling assay [51–53] in HEK293 cells expressing recombinant human NMU1 or NMU2. Examples of concentration-response curves are shown in Figure 2. As a result, we found a lipid-conjugated, truncated analog of NmU, NM4 that possesses high potency at both NMU1 and NMU2, which is similar to that of the native human peptide (hNmU) (Table 1, Figure 1). Notably, NM4 also contains an α, α-disubstituted amino acid (aminoisobutyric acid, Aib) that may confer increased resistance to enzymatic degradation [54].

Figure 2.

Examples of calcium signaling assay dose-response curves for the maximal changes in intracellular [Ca²⁺] ([Ca²⁺]_i) performed in fluo-4-loaded HEK293 cells with stable expression of either human NMU1 (A) or human NMU2 (B).

To test whether newly synthesized NmU analogs possess any in vivo anorectic activity, we compared NM4 and NM7 peptides in a diet-induced obese (DIO) mouse model using previously described protocols [39,41]. Initially we focused on single dose experiments comparing NM4 and NM7 at a dose of 5344 nmoles/kg body weight (Figure 3). NM7 was used as an alternative lead compound because it contains two Aib residues that should confer even greater resistance to degradation within the circulation. This could potentially compensate for lower potency and produce responses equivalent to or better than NM4. Since palmitoylated analogues have limited water solubility, we employed a phospholipid-based, commercially available drug delivery system, PUREBRIGHT^® SL-220 (NOF America Corp., White Plains, NY) that is suitable for delivery of lipidated peptides [44,45]. Direct comparison revealed that in single dose experiment(s) NM4 showed significant anorectic activity (Figure 3A) that was dose-dependent (Figure 3B). However, NM4 contains a dPEG₁₂ linker (40-amino-4,7,10,13,16,19,22,25,28,31, 34,37-dodecaoxotetradecanoic acid, Peptides International, Inc., Louisville, KY), which is particularly expensive and contributes significantly to the cost of peptide production. Therefore we sought a less expensive alternative(s), which resulted in the discovery of the “PEG-free” NM4A analog (Figure 1) that utilized an N-terminal iminodiacetic acid mono-N-palmityl amide (Ida^NHPal). This analog has similar bioactivity to NM4 but is ~10 times less expensive to produce. Subsequent in vivo testing revealed that NM4A is at least 3.5 times more active than NM4 and that its anorectic effect after a single dose (1603 nmoles/kg; 2.2 mg/kg) may last up to ~20 days (Figure 3C). NM4A was characterized further in chronic in vivo experiments. Ad libitum-fed male DIO C57BL/6 mice, which were maintained on a high fat diet (Cat# D12492, Research Diets, Inc., New Brunswick, NJ), were treated subcutaneously (s.c.) with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) either every 2, 4 or 7 days for a period of 21 days. As shown in Figure 4A, NM4A exerts potent anorectic effects when injected every 2 or every 4 days, with less frequent dosing showing limited activity in this chronic administration regime. Notably s.c. administration of NM4A at the abovementioned doses caused initial gastric emptying within 24 hours of administration but such effects subsided with time. To indicate whether NM4A would be a useful adjunct in obese subjects who are dieting, we tested its efficiency in a once weekly regimen at a dose of 2.2 mg/kg. In this case, ad libitum fed male DIO C57BL/6 mice were kept on either the high fat diet (FD) or regular (normal) diet (ND) during the administration of NM4A. The data suggest that NM4A is beneficial under such circumstances, showing an additional decrease in body weight (Figure 4B).

Figure 3.

Single dose in vivo experiments performed in the DIO mouse model. (A) Comparison of anorectic effects of NM4 and NM7. (B) NM4 dose-response experiments. (C) Comparison of NM4 and NM4A anorectic activity.

Figure 4.

Chronic in vivo experiments performed in the DIO mouse model.. (A) Ad libitum fed male DIO C57BL/6 mice were treated with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 2, 4 or 7 days for 21 days. (B) Ad libitum fed male DIO C57BL/6 mice were treated with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) once a week for 21 days. After the initial treatment animals were kept either on high-fat diet (FD) or normal diet (ND). (C) Ad libitum fed male DIO C57BL/6 mice were treated with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) or NM4-C₁₆ (2.6 mg/kg; 1603 nmoles/kg) every 7 days for 21 days.

Considering current trends in obesity/diabetes treatment [55], the development of long-acting analogs of NmU is highly desirable. Such analogs could be used in combination therapies with, for example, incretin-based therapeutic(s) such as semaglutide [56–58], a long-acting glucagon-like peptide 1 (GLP-1) analog that possesses a plasma half-life of 160 h. Since NM4A showed limited activity in a once-weekly regimen we synthesized double-lipidated analogs that should theoretically have improved pharmacokinetic properties. For example, plasma half-life should be increased as a result of the increased possibility/strength of hydrophobic interactions with abundant plasma proteins (albumin, HDL, etc.) [59]. In addition, lipid-conjugation may improve bio-activity by increasing the local concentration of the analog(s) in the hydrophobic environment of lipid rafts within cell plasma membranes [60–62] that are enriched in various GPCRs [63–66]. Therefore, using a reductive-alkylation approach, we synthesized two analogs: NM4A-C₁₆ and NM4-C₁₆, with varying lipidation points/geometry (Figure S2). In vitro testing revealed that only NM4-C₁₆ with a symmetrical lipidation motif that is removed from main body of the NM-peptide, possesses potent bioactivity (albeit reduced compared to NM4A) (Table 1). To indicate whether this approach had yielded an analog that could be used in a once-weekly injection regime, we performed pharmacokinetic studies of NM4A and NM4-C₁₆. The double-lipidated analog, NM4-C₁₆, did indeed have an extended plasma half-life (t_1/2=28.2±1.0 h) compared to the single-lipidated NM4A (t_1/2=24.0±1.0 h) although the effects of double-lipidation were limited (Figure 5). These results, as well as the relatively short plasma half-life of both, PEG- and HSA-NmU conjugates reported elsewhere [40,41] (t_1/2≈11–54 h, depending on species) suggested that a key factor for the in vivo activity of NmU analogs/derivatives may be their resistance to enzymatic proteolysis. Alternatively, loss of their activity may involve partial hydrolysis of the N-terminal N⁷-amide to α- or β-monocarboxylic acid residues (X-Asn-OH or X-Asp-NH₂) via an aminosuccinimide intermediate [22]. Therefore we performed plasma stability studies for both NM4A and NM4-C₁₆ analogs using full length hNmU as a comparator [67–69]. Results (Figure 5C) indicated that both analogs have significant resistance to mouse plasma driven enzymatic degradation, with NM4-C₁₆ being more stable than NM4A. Notably, under the same experimental conditions parental hNmU is quickly degraded falling to less than 1% of initial content within 4 h. Since plasma stability studies do not fully recapitulate physiological conditions (e.g. lack of membrane(s) bound enzymes, different ratios and distribution, etc.) the predictive value of such experiments is rather limited. Nonetheless, the plasma stability of NM4A at t_1/2 (24 h) is 79.4±1.8% and plasma stability of NM4-C₁₆ at its t_1/2 (28 h) is 98.7±1.9% suggesting that enzymatic resistance may play limited role in performed animal experiments, as at respective t_1/2 time points both peptides remain largely intact. Even with a prolonged time of exposure (144 h = 6 days) amounts of both peptides remain significant reaching 39.9±0.4% for NM4A and 60.8±0.4% for NM4-C₁₆.

Figure 5.

Pharmacokinetics and plasma stability of NM4A and NM4-C₁₆ analogs. For pharmacokinetic studies animals were individually dosed with 10 mg/kg of (A) NM4A or (B) NM4-C₁₆. Plasma stability studies (C) were performed in freshly prepared mouse plasma at 25 μM final concentration(s) of tested analogs.

Further modifications of NM4-C₁₆ are limited due to the high specificity of NMU1 and NMU2 [6,28]. Indeed, our attempts to increase proteolytic resistance of truncated NmU analogs proved largely unsuccessful, with various retro-inverso-, α-carbon-methylated, N-methylated and cyclic analogs being in most cases virtually inactive (Table 1). Nonetheless, data obtained for the analogs NM8-16 suggest that substitution of Pro for more rigid and bulky residues, e.g. Oic or Tic, has an undesirable effect (potency: NM8>NM9>NM10). Moreover, comparison between analogs NM8-10 (linear peptides), NM11-13 (cyclic peptides) and NM14-16 (cyclic/lipidated peptides) also suggests that cyclization is not a viable option as it renders analogs NM11-13 inactive. Interestingly, introduction of a lipid moiety to the cyclic analogs (NM14-16) seems, at least partially, to restore bioactivity. These results are generally in line with published observations that underline the importance of P⁵ and N⁷-amide residues for agonistic activity at NMU receptors [21–31]. Considering the inactivity of our retro-inverso-analogs (NM3D1-3) and α-carbon-/N-methylated-variants (NM31-NM35), as well as the limited activity of cyclic derivatives (NM14) and derivatives containing D-amino acids [29,31], further derivatization and/or modifications of NmU for pharmaceutical purposes appears to be extremely difficult since even small changes in the structure may lead to significant loss of bioactivity. Nonetheless, a viable modification seems to be the introduction of a small α, α-disubstituted amino acid(s) (Aib) at the flanking positions of the native active sequence (e.g. FLFRPRN). Notably bis-substituted analogs (e.g. X-Aib-FLFRPRN-Aib) show decreased potency in in vitro assays (NM5<NM2 & NM6<NM3) and very limited activity in vivo (NM7, Figure 3A). Lipidation itself does not appear to have detrimental effects, provided its position is significantly removed from the active portion of the analog(s) (NM4 versus NM2). Although the impact of lipidation on full-length human NmU analogs seem to be more diverse depending on their individual structure [31].

Despite limited in vitro activity of the double-lipidated NM4-C₁₆, this analog was tested in vivo using a once weekly s.c. injection regime at 2.6 mg/kg, which is equimolar to the dose of NM4 used previously (i.e. 1603 nmoles/kg). As shown in Figure 5C, once-weekly administration of NM4-C₁₆ resulted in sustained weight loss in the DIO mouse model. Moreover, in this case we observed very limited gastric emptying in experimental animals and their hematological parameters/blood cell counts were within “normal” range (Table 2). In this particular case, we believe that observed discrepancies between in vitro an in vivo activity may be attributed to limited water solubility of NM4-C₁₆ which limits effective concentrations of the compound in in vitro settings.

Table 2.

Hematological parameters of animals treated with NM4-C₁₆. Ad libitum fed male DIO C57BL/6 mice (high-fat diet) were treated with NM4-C₁₆ (2.6 mg/kg; 1603 nmoles/kg) every 7 days for 21 days. Blood was collected by cardiac puncture and cell counts performed using a HemaVet 950FS Hematology Analyzer (Drew Scientific Inc., Dallas, TX). Control animals received vehicle alone.

Cells	Count/mL
	Treated	Control
White Blood Cells	5.73 ± 0.49×106	7.35 ± 1.21×106
Neutrophils	1.02 ± 0.06×106	1.30 ± 0.21×106
Lymphocytes	4.56 ± 0.43×106	4.82 ± 0.68×106
Monocytes	1.46 ± 0.14×105	1.07 ± 0.31×105
Red Blood Cells	9.49 ± 0.31×109	9.17 ± 0.17×109
Platelets	6.95 ± 0.23×108	5.97 ± 0.32×108

For three groups of animals treated every 2, 4 or 7 days we also evaluated abdominal fat content (Figure 6A). These data indicate that chronic treatment with either, NM4A or NM4-C₁₆ decreases abdominal fat content from 8.1±0.4% (control) to 5.3±0.2%, 4.9±0.3% and 4.5±0.4% for animals treated every 2, 4 and 7 days respectively.

Figure 6.

(A) Abdominal fat of chronically treated animals. Ad libitum fed male DIO C57BL/6 mice were treated with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 2 days or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 4 days or ^(*) NM4-C₁₆ (2.6 mg/kg; 1603 nmoles/kg) every 7 days for 21 days. Subsequently, their abdominal fat was harvested and weighed. Results are expressed as abdominal fat as a % of total body weight. (B) Distribution of NM4A in NM4A chronically treated animals (2.2 mg/kg; every 2 days for 28 days). Subsequently their organs were harvested and analyzed using MDS Sciex QSTAR XL Hybrid Quadrupole Time-of-Flight LC/MS (Applied Biosystems) with methylated NM4A (NM4A-Me) as an internal standard. (C) LPA levels in the plasma of chronically treated animals with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 2 days or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 4 days or ^(*)NM4-C₁₆ (2.6 mg/kg; 1603 nmoles/kg) every 7 days for 21 days. (18:1)-1-oleoyl-2-hydroxy-sn-glycero-3-phosphatidic acid, (18:2)-1-linoleoyl-2-hydroxy-sn-glycero-3-phosphatidic acid, (20:4)-1-arachidonoyl-2-hydroxy-sn-glycero-3-phosphatidic acid.

To examine the body distribution of NM-analogs we performed limited ADME studies in animals that underwent chronic treatment with NM4A every 2 days for 28 days (2.2 mg/kg; 1603 nmoles/kg). As shown in Figure 6B, the highest content of NM4A was detected in the plasma of experimental animals, with significant amounts also accumulating in the liver, lungs and kidneys. Notably NM4A does not appear to enter the brain, suggesting that its biological effects are transduced via NMU receptors present in the periphery.

The therapeutic use of NmU analogs in obesity and diabetes will require prolonged treatment regimes. Since NmU has cardiovascular effects [70,71] as well as being indicated in cancer [72–74], it is of great interest whether prolonged exposure to our analogs may have related side effects. To assess such possible effects we measured levels of lysophosphatidic acid (LPA, (20:4)-1-arachidonoyl-2-hydroxy-sn-glycero-3-phosphatidic acid; (18:2)-1-linoleoyl-2-hydroxy-sn-glycero-3-phosphatidic acid; and (18:1)-1-oleoyl-2-hydroxy-sn-glycero-3-phosphatidic acid) in the plasma of animals treated frequently (1603 nmoles/kg; every 2 or 4 or 7 days for 21 days) with NM4A or NM4-C₁₆ using a previously described LC-ESI-MS/MS-based method [75–78]. LPA is a ubiquitously present bioactive phospholipid which plays important roles in diverse cellular processes such as cell migration, proliferation and differentiation by acting on its cognate GPCRs [79]. The LPA signaling network is prominent in atherosclerosis development [80–84], cancer initiation, progression and metastasis [85–96] and has also been linked to insulin resistance [97–99] making it well suited for assessment of possible side effects of our therapeutics. As shown in Figure 6C, prolonged treatment of DIO mice with NM4A does not increase plasma levels of LPA suggesting that therapy should have no negative effects on obese patients that are prone to atherosclerosis and diabetes development. Since LPA is also indicated in cancer, these results also suggest that use in obese cancer patients or people who are not aware they have cancer, should be relatively safe, which may be of concern due to the connection of NmU with breast and pancreatic cancers [72–74]. In the case of NM4-C₁₆ slightly elevated levels of (20:4) were observed which may be of concern since this particular lipid was shown to enhance progression of atherosclerosis in animal model [82].

To examine whether lipidation and the position of the lipid addition has an impact on the secondary structure of the NM-analogs, we performed FTIR [100] studies of selected compounds that are listed in Supplemental Material (Figure S2 & S3). Results are summarized in Table 3. An examination of the secondary structure of the full length hNmU by infrared spectrometry indicated that the peptide has a broad absorbance spectrum in the amide I conformation band centered at 1657 cm⁻¹ (Figure S4). Deconvolution of this broad absorbance revealed strong contributions from α-helix and loop-turn components (Table 3). The turn components were adjacent to the α-helical absorbance suggesting that they represented strong 310-helix (type III turn propensity) conformations typical of peptides with Aib residues [101]. The well-defined β-sheet peak centered at 1629 cm⁻¹ was also present. Shorter versions of the parent sequence either had a significant disordered structure as seen with NM1 and its palmitoylated derivative NM3 or significant β-sheet components as seen with the spectrum of NM2 and NM4C13 (Figure S4; Table 3). With the exception of NM4C13, all NM4 N-terminal derivatives had a strong helix–turn conformation suggesting that the addition of N-terminal lipid, PEG or related components helped stabilize the helix and turn propensity of the truncated hNMU sequence. These results further suggest that the N-terminal lipid derivatization not only stabilizes the short peptide structure but may provide an additional hydrophobic anchoring of the active conformation to the active site.

Table 3.

Analysis of peptides secondary structure by FTIR spectroscopy.

Peptide*	% Conformation
	α-helix	β-sheet	loop-turn	disordered
hNMU	30.36	26.20	24.57	18.87
NM1	30.35	15.95	21.79	31.91
NM2	26.36	31.56	27.57	11.51
NM3	35.24	15.95	21.79	31.91
NM4	30.40	16.95	31.69	20.96
Ac-NM2	30.97	25.95	34.66	20.96
NM4C13	26.46	19.36	34.62	19.56
NM4A	29.84	20.75	21.69	27.72
NM4A-COOH	27.77	16.72	25.92	29.59
NM4-Me	34.05	15.27	33.08	17.60
NM4A-Nic	29.47	18.34	29.27	22.92
NM4A-C16	45.93	20.49	20.17	13.41
NM4-C16	30.31	16.97	27.72	25.00

^*self-films of peptides (1.2 mM) hydrated with HFIP:D₂O 10 mM buffer pD=7.4 (v:v, 4:6) were analyzed for secondary conformation based on secondary structural analysis using spectral deconvolution and curve fitting.

3. Discussion

In recent years obesity became a global health problem, and the development of anti-obesity therapeutics attracted much attention [55]. Our recent study aimed at the development of short, lipidated peptide analogs, derived from NmU, that could be used as such therapeutics. Using various peptides’-stabilization-protocols, including: cyclization, lipidation, introduction of α, α-disubstituted amino acids, a retro-inverso-approach, and a combination of these, we synthesized a group of truncated NmU analogs, that underwent an extensive in vitro and in vivo testing. As a result, three potent truncated/lipidated agonists on NmU were found: NM4, NM4A and NM4-C₁₆. Unlike previously described PEG-, and HSA-conjugates that require an NmU-carrier conjugation step, these analogs can be efficiently synthesized in large quantities and lipid conjugated using SPPS technology giving the final product with a high yield. In addition, they possess short, 8 amino acids long core sequence, which is approximately 1/3 of native hNmU. Taking into account the molecular weight of reported conjugates [40,41] our analogs are approximately 40 times smaller allowing for delivery of larger quantities of active ingredient (per mole count) in relatively smaller doses. Moreover, use of smaller, palmitoylated peptide(s) may be advantageous over conjugates with macromolecular carriers due to possible side effects that may be associated with the latter. Namely, conjugates with HSA may lead to generation of an antibody-based immune response. Hypersensitivity to macromolecular-PEGs has also been reported [102] triggering, for example, withdrawal of the PEGylated erythropoietic agent Omontys^® from the market. Importantly, our double-lipidated analog NM4-C₁₆, was active in vivo in a once weekly s.c. injection regimen, which suggests it could be used in combination therapies with a long-acting glucagon-like peptide 1 (GLP-1) analog(s) during treatment of diabetic patients. Since administration of GLP-1 derivatives also leads to weight loss, it is reasonable to assume possible synergistic effects. However, as NmU is implicated in both insulin secretion [103,104], and glucose homeostasis [39], successful application of such combination therapy will certainly require an extensive further experimentation. Generally, our results demonstrated utility of lipidation approach in modification of short NmU derivatives producing physiologically active derivatives. As the core sequence of these analogs was only minimally altered with proteolytic-resistance-inducing substituents (Aib residue at N-terminus), lipid-conjugation appears to be particularly well suited for the synthesis of therapeutically relevant NmU agonists. Nonetheless, further derivatization and/or modification of NmU for pharmaceutical applications appears to be particularly difficult since even small changes in the structure generally lead to significant loss of bioactivity. Moreover, lingering problem of receptor selectivity (NMU1 versus NMU2) remains largely unsolved, at least in the context of pharmaceutical applications. However, synthesis of NMU1- and NMU2-receptor-specific agonists was recently reported [28], which certainly stimulate further studies in this promising therapeutic area.

4. Conclusions

New family of truncated/lipidated agonists of neuromedin U was synthesized, characterized, and tested both, in vitro and in vivo using a DIO mouse model. Administration of selected peptides in various regimens resulted in significant weight loss in experimental animals. We believe that these analogs may serve as leads for the development of anti-obesity therapeutics.

5. Experimental section

5.1. Peptide synthesis

All peptides were synthesized as C-terminal cysteamine-amides by the solid phase method using a CEM Liberty automatic microwave peptide synthesizer (CEM Corporation Inc., Matthews, NC), applying 9-fluorenylmethyloxycarbonyl (Fmoc) chemistry and commercially available amino acid derivatives and reagents (EMD Biosciences, San Diego, CA and Chem-Impex International, Inc., Wood Dale, IL). Cysteamine 2-Chlorotrityl Resin (EMD Biosciences, San Diego, CA) was used as a solid support. Peptides were cleaved from resin using modified reagent K (TFA 94% (v/v); phenol, 2% (w/v); water, 2% (v/v); TIS, 1% (v/v); EDT, 1% (v/v); 2 h) and precipitated by addition of ice-cold diethyl ether. Subsequently, peptides were purified by preparative reverse-phase high performance liquid chromatography (RP-HPLC) and their purity evaluated by matrix-assisted laser desorption ionization spectrometry (MALDI-MS) as well as analytical RP-HPLC.

5.2. Synthesis of NM4A-C₁₆, NM4-C₁₆ and NM4A-Me by reductive alkylation

The analogs NM4A-C₁₆, NM4-C₁₆ and NM4A-Me (for FTIR and PK studies only) were synthesized from NM4A, Ahx-Aib-FLFRPRN-amide and NM4A respectively. Peptides were dissolved in a mixture of 1,4-dioxane, methanol and water (5:4:1) and mixed with proper carbonyl compound (50 eq, 30 min). For analogs NM4A-C₁₆ and NM4-C₁₆ commercially available 1-hexadecanal was used (Cayman Chemical Company, Ann Arbor, MI), and in the case of NM4A-Me formalin was employed. Subsequently, acetic acid was added (100 eq) and the reaction mixture was placed in an ice-bath and mixed for an additional 20 min (magnetic stirrer). Subsequently freshly prepared water solution of NaBH₃CN (100 eq) was added dropwise with vigorous mixing and afterwards the reaction mixture was agitated for an additional 3 hours. The reaction mixture was then further acidified with acetic acid (500 eq), diluted with water (1:1) and freeze-dried. Obtained solid residue(s) were purified by preparative RP-HPLC and their purity evaluated by MALDI-MS as well as analytical RP-HPLC.

5.3. Synthesis of cyclic analogs NM11-16

Cyclization of the linear analogs NM8-NM10 was performed in a 50% solution of DMSO in DMF using 1,3-bis(bromomethyl)benzene (NM11-NM13) or 1-(palmityl-S-methyl)-3,5-bis(bromomethyl)-benzene (NM14-NM16). Briefly, peptides (2 eq.) were dissolved at final concentrations of 5 mg/ml. Subsequently anhydrous cesium carbonate (Cs₂CO₃, 20 eq.), tetrabutylammonium iodide, (TBAI, 4 eq.) and proper bis(bromomethyl)-benzene-derivative (1 eq) were added. The solution was vigorously mixed on a magnetic stirrer and the progress of the reaction monitored by analytical RP-HPLC. Subsequently peptides were purified and characterized as described in the peptide synthesis section.

5.4. Analytical RP-HPLC

Analytical RP-HPLC was performed on a Varian ProStar 210 HPLC system equipped with ProStar 325 Dual Wavelength UV-Vis detector with the wavelengths set at 220 nm and 280 nm (Varian Inc., Palo Alto, CA). Mobile phases consisted of solvent A, 0.1% TFA in water, and solvent B, 0.1% TFA in acetonitrile. Analyses of peptides were performed with either an analytical reversed-phase C₄ XBridge^™ BEH300 column, 4.6×150 mm, 3.5 μm (Waters, Milford, MA), or an analytical reversed-phase C₁₈ SymmetryShield^™ column, 4.6×250 mm, 5 μm (Waters, Milford, MA) applying linear gradient of solvent B from 0 to 100% over 100 min (flow rate: 1 ml/min).

5.5. Cellular calcium signaling assay

HEK293 cells expressing either the neuromedin U 1 receptor or neuromedin U 2 receptor (HEK-NMU1 and HEK-NMU2 respectively)56 were grown to approximately 90% confluence in ELISA strip plates (96-well format) pre-coated with poly-D-lysine (0.1% w/v) and loaded for 45 min at 37 °C with 2 μM fluo-4-AM in Krebs’-HEPES buffer with bovine serum albumin (KHB-BSA; composition: 10 mM HEPES; 4.2 mM NaHCO₃; 11.7 mM D-glucose; 1.18 mM MgSO₄·7H₂O; 1.18 mM KH₂PO₄; 4.69 mM KCl; 118 mM NaCl; 1.3 mM CaCl₂·2H₂O; 0.1% (w/v) BSA; pH 7.4). Monolayers were then washed and equilibrated for 5 min at 37 °C in 100 μl KHB-BSA for subsequent recording of fluorescence as an index of intracellular [Ca²⁺] ([Ca²⁺]_i) using a microplate reader (NOVOstar; BMG LABTECH, Aylesbury, U.K.). Briefly, 20 μl of KHB-BSA, with or without ligand, was added into the well (200 μL/s) and fluorescence determined at 0.5 s intervals by excitation at 485 nm and collection of emitted light at 520 nm. Changes in fluorescence above basal levels (before ligand addition) were determined. When required, [Ca²⁺]_i was calculated using the formula: [Ca²⁺]_i = K_d (F−F_min)/(F_max−F), with the K_d of fluo-4 taken as 350 nM [53]. F_max was obtained by removal of buffer and addition of KHB-BSA buffer containing 4 mM [Ca²⁺] and ionomycin (2 μM) to representative wells and the fluorescence measured for 10 min. F_min was then derived by replacing buffer with Ca²⁺-free KHB-BSA buffer containing 2 mM EGTA and fluorescence measured for 10 min [51]. For the determination of concentration-response relationships, the maximal change in [Ca²⁺]_i was calculated following ligand addition. Concentration-response curves were then fitted using a four-parameter logistic equation to determine EC₅₀ values (GraphPad Software Inc., CA).

5.6. Animal experiments

All animal experiments were approved by the UCLA Animal Care and Use Committee (ARC#1999-173-23) and conformed to local and national guidelines. Ad libitum fed male DIO C57BL/6 mice (n=8 per group) kept on D12492, a high-fat diet composed of 60% Kcal from fat (Research Diets, Inc., New Brunswick, NJ) were weighed and individually dosed (s.c., 30 min prior to the onset of the dark phase of the light cycle) with either vehicle or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 2 days or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 4 days or NM4A (2.2 mg/kg; 1603 nmoles/kg) every 7 days or NM4-C₁₆ (2.6 mg/kg; 1603 nmoles/kg) every 7 days for 21 days and their body weight were measured daily. Due to limited water solubility, the lipidated analogs NM4A and NM4-C₁₆ were formulated in a phospholipid-based, commercially available drug delivery system, PURE-BRIGHT^® SL-220 (NOF America Corp., White Plains, NY). Briefly, the calculated amount of the desired peptide (1 eq) and 10 equivalents of PUREBRIGHT^® SL-220 (1 mg : 10 mg) were solubilized in a minimal volume of 90% absolute ethyl alcohol in water. Subsequently the solution was aliquoted into Eppendorf tubes and freeze-dried. The solid residue was re-solubilized in an appropriate amount of PBS directly before use by vortexing.

5.6.1. Body distribution analysis

Ad libitum fed male DIO C57BL/6 mice that underwent chronic treatment with NM4A (2.2 mg/kg; 1603 nmoles/kg; dosing every 2 days for 28 days) were sacrificed and their organs/tissues harvested and weighted. Subsequently 4 volumes of a DMSO/ACN mixture (1:1) containing 0.1% of TFA, were added and samples homogenized using a rod homogenizer (max. speed for 30 s). The homogenate(s) were transferred to 1.5 mL Eppendorf tubes and centrifuged for 10 min at 13000 rpm. Obtained supernatant(s) were analyzed the Agilent 6460 Triple Quadrupole LC/MS System (Agilent Technologies, Santa Clara, CA) with methylated NM4A (NM4A-Me) as an internal standard.

5.6.2. Pharmacokinetic (PK) studies

C57BL/6 mice were weighted and individually dosed with either NM4A or NM4-C₁₆ at 10 mg/kg dose. Subsequently small samples of blood were collected at the indicated time-points and centrifuged (3000 rpm/10 min). Obtained plasma samples were transferred into the 0.5 mL centrifuge tubes and immediately diluted with 4 volumes of a DMSO/ACN mixture (1:1) containing 0.1% of TFA. Subsequently samples were centrifuged at 13000 rpm for 10 min and obtained supernatants analyzed using the Agilent 6460 Triple Quadrupole LC/MS System (Agilent Technologies, Santa Clara, CA) with methylated NM4A (NM4A-Me) as an internal standard.

5.7. Plasma stability studies

Tested analog(s) (10 mM stock solutions in DMSO) were added to freshly prepared mouse plasma (1 μL per 400 μL of plasma, c=25 μM) and incubated at 37 °C. Subsequently small samples of plasma (10 μL) were collected at the indicated time-points and immediately diluted with 200 μL of a DMSO/ACN mixture (1:1) containing 0.1% of TFA. Samples were then centrifuged at 13000 rpm for 10 min and obtained supernatants analyzed using the Agilent 6460 Triple Quadrupole LC/MS System (Agilent Technologies, Santa Clara, CA) with methylated NM4A (NM4A-Me) as an internal standard.

5.8. LPA measurements

LPA was determined by LC-MS/MS as previously described [75–77]. Briefly, plasma samples (100 μL) were extracted using 3cc, 60 mg, Waters Oasis HLB solid phase extraction cartridges. The extracted samples were evaporated under inert gas to dryness using low heat (30–37 °C), and reconstituted with 75 μL of methanol with vortexing. The extracts were transferred to Eppendorf tubes and centrifuged to remove debris. The supernatants were analyzed using the Agilent 6460 Triple Quadrupole LC/MS System (Agilent Technologies, Santa Clara, CA) and quantified using Agilent Mass Hunter QQQ Quantitative Analysis software.

5.9. Determination of peptide secondary structure by FTIR

Since the NM peptides used in this study have a lipo-peptide like character, samples were measured as self-films dried onto a germanium ATR (Attenuated Total Reflectance) sampling crystal. Infrared spectra were recorded at 25 °C using a Bruker Vector 22 FTIR spectrometer equipped with a deuterated triglycine sulfate (DTGS) detector, averaged over 256 scans at a gain of 4 and a resolution of 2 cm⁻¹. FTIR spectra of the peptide samples were obtained from peptide self-films prepared by dissolving peptides in hexafluoroisopropanol and then air drying the solution onto 50×20×2 mm 45° ATR crystals fitted for the Bruker spectrometer (Pike Technologies, Madison, WI). The dried peptide self-film was then hydrated by exposing the sample to nitrogen gas saturated with vapor from deuterated 10 mM sodium phosphate buffer (pH=7.4) for one hour prior to spectral acquisition. The relative proportions of α-helix, turn-loop, β-sheet, and disordered conformations of solution and multilayer IR spectra were determined by Fourier self-deconvolution for band narrowing and area calculations of component peaks of the FTIR spectra using curve-fitting software supplied by Galactic Software (GRAMS/AI, version 8.0; Thermo Electron Corp., Waltham, MA). The frequency limits for the different structures were: α-helix (1662–1645 cm⁻¹), β-sheet (1637–1613 cm⁻¹ and 1710–1682 cm⁻¹), turns (1682–1662 cm⁻¹), and disordered or random (1650–1637 cm⁻¹) [100].

Supplementary Material

supplement

Figure S1. An example of analytical RP-HPLC profile and corresponding MS-spectra obtained for analog NM4-C₁₆.

Figure S2. Structures of neuromedin analogs tested in FTIR experiments-Series 1.

Figure S3. Structures of neuromedin analogs tested in FTIR experiments-Series 2.

Figure S4. FTIR spectra of self-films of Series 1 & 2 peptides hydrated with HFIP : D₂O 10 mM buffer pD=7.4 (v:v, 4:6).

Highlights.

• Truncated analogues of neuromedin U are active in vivo.

• Lipidated derivatives of NmU are viable drug candidates for obesity treatment.

• Position of lipidation influences affinity of NmU derivatives to NMU1 and NMU2 receptors.

• The pharmacokinetic properties of selected lipidated NmU derivatives were studied in murine model.

Appendix A. Supplementary data

Supplementary data (representative analytical RP-HPLC profile and corresponding MALDI-MS spectra, structures of NM analogs tested in FTIR experiments and corresponding FTIR spectra) associated with this article can be found, in the online version, at…..