Liquid chromatography/mass spectrometry based detection and semi‐quantitative analysis of INSL5 in human and murine tissues

Rationale Insulin‐like peptide 5 (INSL5) is a hormone produced by enteroendocrine L‐cells in the colon that has recently been implicated in the control of metabolic homeostasis. However, research into its physiology has been hindered by the reported unreliability of commercially available immunoassays and additional detection assays would benefit this emerging field. Methods Peptides from purified murine L‐cells and homogenates from both human and mouse colonic tissues were extracted by precipitating larger proteins with acetonitrile. Untargeted liquid chromatography/tandem mass spectrometry (LC/MS/MS) analyses, followed by database searching, were used to detect and identify various INSL5 gene derived peptides and characterise their precise sequence. A similar approach was developed to quantify INSL5 levels in primary intestinal culture supernatants after purification and concentration by solid‐phase extraction. Results Mass spectral analysis of purified enteroendocrine cells and tissue homogenates identified the exact sequence of A and B chains of INSL5 endogenously expressed in L‐cells. Differences in the endogenously processed peptide and the Swissprot database entry were observed for murine INSL5, whereas the human sequence matched previous predictions from heterologous expression experiments. INSL5 was detected in the supernatant of human and mouse primary colonic cultures and concentrations increased after treatment with a known L‐cell stimulus. Conclusions The first LC/MS/MS‐based method capable of the detection and semi‐quantitative analysis of endogenous INSL5 using MS‐based techniques has been demonstrated. The methodology will enable the identification of stimulants for INSL5 secretion from murine and human primary colonic epithelial cultures.

is expressed at highest levels in the colon of both mice and humans, with lower expression found for example in thymus and testis 1,2 and within the colon it is expressed in enteroendocrine L-cells, 3 which also secrete glucagon-like peptide-1 (GLP-1) 4 and peptideYY (PYY). 4 Liu and collaborators showed that INSL5 stimulates the G-protein coupled receptor RXFP4 (GPR100 or GPCR142) but no precise physiological function of the peptide was initially identified. 5 Since then INSL5 has been implicated in the regulation of metabolism, either by affecting glucose tolerance through an impact on insulin production, 6 insulin the stimulation of food-intake. 3 Consistent with an orexigenic function, Grosse et al 3 reported a fall in INSL5 plasma levels after refeeding of starved mice; however, they observed poor reproducibility and batch dependence of several commercially available immunoassays. The detected plasma concentrations varied widely, 3 but generally were in the range of 1-100 pg/mL, whereas a recent report detected 100-fold higher INSL5 levels in human serum. 9 Importantly, however, with commercial immunoassays we were unable to detect INSL5 either in supernatants or lysates of primary colonic epithelial cultures that we routinely use to investigate secretion of GLP-1 and PYY from L-cells. As we were able to detect INSL5 in these cultures by immunohistochemistry, 3 we aimed to develop an independent assay suitable for the investigation of stimulus secretion coupling in these in vitro systems.
Over the last two decades, mass spectrometry (in combination with electrospray ionization) has become a powerful tool in the analysis of peptides, both for quantitative analysis and for identification of potential new functional peptides. Studies have included the analysis of primary islet cell cultures where significant numbers of glucagon gene products were identified. 10 More targeted mass spectrometry techniques using triple quadrupole systems have been employed to monitor peptides such as GLP-1 7-36 and 9-36 in established colorectal cell lines. 11 Furthermore, mass spectrometric approaches have proven to be sensitive enough to detect some gut hormone peptides at physiological (usually low pM or pg/mL) levels in circulating plasma. 12,13 Mass spectrometry has also been successfully applied to the detection and quantitation of peptides from the insulin/relaxin family. In contrast to most immunoassays, mass spectrometry can readily distinguish between peptides with only subtle differences as for example endogenous human insulin and the closely related Humalog, which is isobaric and differs only by a transposition of the proline and lysine residues at the C-terminal end of the A-chain. 14 This study also demonstrated that mass spectrometric approaches can reach lower limits of quantitation in the region of 50 pg/mL in plasma for most of the insulin analogues. We thus investigated the possibility of using a MS-based assay for INSL5 that would also allow the characterisation of the exact peptide sequence of stored and secreted products derived from the prepropeptide.

| Peptide extraction from FACS-sorted murine L-cells
Single cell suspensions of colonic epithelial cells were prepared with slight modifications to a previously described procedure. 15 16 were harvested and washed with phosphate-buffered saline (PBS). The external muscle layer was removed and tissue was incubated in PBS with 15 mM EDTA and 0.5mM dithiothreitol (DTT) for 10 min, transferred into cold PBS with Y27632 (10 μM) and shaken for 30 s to detach crypts. This was repeated five times and crypts and cell clumps detached from the tissue were centrifuged (300 g, 5min) after each step.
Pooled sediments were digested into single cells in 0.25% trypsin-EDTA Single live L-cells were isolated based on forward and side scatter and Venus, DAPI and Draq5 fluorescence within 4 h using a FACSJazz (BD Biosciences, Singapore) at the Cambridge NIHR BRC Cell Phenotyping Hub. Peptides were extracted using a method adapted from a validated approach for enriching peptides from plasma, [17][18][19] where a total of 6000-8000 positive cells were sorted and lysed directly into 800 μL of 80% ACN in water (v/v) in a Protein Lobind Eppendorf tube. Samples were centrifuged for 5 min at 10,000 g to pellet the proteins and the supernatant was transferred to a fresh tube and evaporated to dryness in an Eppendorf 5301 vacuum concentrator. Reduction and alkylation of disulphide bonds was performed (as described below) prior to LC/ MS analysis, and 10 μL was injected onto the system.

| Reduction and alkylation
The sorted cell peptide residue was reconstituted into 25 μL of 50 mM ammonium bicarbonate containing 10 mM DTT and heated at 60°C for 60 min. Then 5 μL of 100 mM iodoacetamide (IAA) in 50 mM ammonium bicarbonate were mixed in and the samples were stored in the dark for 30 min at room temperature and subsequently exposed to light for at least 30 min. 5 μL of 1% formic acid in water (v/v) was added prior to injection onto the LC/MS system. prior to injection onto the LC/MS system. To confirm the sequence of the A and B chains, the SPE eluent (50 μL) was dried under oxygen-free nitrogen (heated to 40°C) on a SPE Dry evaporator system (Biotage, Upsalla, Sweden) prior to reduction and alkylation.

| Nano LC/MS analysis of sorted cells and tissue homogenates
Peptide extracts were analysed using a Thermo Fisher Ultimate 3000 nano-LC system coupled to a Q Exactive Plus Orbitrap mass spectrometer

| Endogenous peptide identification
The acquired LC/MS files were searched using Peaks 8.0 software (Waterloo, ON, Canada) against the mouse and human Swissprot databases (Downloaded on 06-May-2016). A no-digest setting was used, which enabled peptides of up to 65 amino acids in length to be matched, and precursor and product ion tolerances were set at 10 ppm and 0.05 Da, respectively. A fixed post-translational modification of carbamidomethylation was applied to cysteine residues, whilst variable modifications included methionine oxidation, N-terminal pyro-glutamate, N-terminal acetylation and C-terminal amidation. A false discovery rate value of 1% was used to filter the results, with a minimum of 1 unique peptide also required.

| Precision and accuracy of the SPE approach
An attempt was made to develop a quantitative methodology for the intact murine INSL5 peptide using a commercially sourced murine INSL5 peptide. However, the characterisation of the peptide demonstrated it was different from the endogenous peptide, and therefore could not be used to generate a fully quantitative methodology.
The peptide was however used to assess the sensitivity, precision and accuracy of the SPE approach for a very closely related peptide   Figure 3C shows the [M+5H] 5+ charge state of the human INSL5, which suggests that the peptide had a molecular weight that was calculated as 5041.43 Da, which is very close to the  The mass spectrometric analysis of the Phoenix peptide suggests that the molecule was as sold; however, as mentioned previously, the peptide has an N-terminal serine residue on the B chain, which differs from both the Swissprot specified sequence and the detected endogenous peptide. The Phoenix peptide standard did in fact contain a small amount of the endogenous peptide; however, this appeared to be present as an impurity, as its peak area was approximately 1.5% of that attributed to the Phoenix INSL5 with the additional N-terminal serine residue. The difference in amino acid sequence between these two versions of murine INSL5 might not affect immunoassay-based quantitation methods if the antibodies epitope is elsewhere in the INSL5 three-dimensional structure. However, the addition of the serine residue means that mass spectrometric analysis approaches are compromised due to the significant change in the mass of the peptide.

| LC/MS analysis of intact endogenous and synthetic INSL5 peptides
One potential approach to use the Phoenix peptide for quantitative studies is to reduce and alkylate the peptide, separating the two chains and analysing the A chain only, as this is identical to the endogenous version. However, using this approach would result in the loss of structural information, where any cleavages or modifications to the B chain would be lost if only the A chain is monitored.

| Assessment of the precision and accuracy of the SPE method
The precision and accuracy of the SPE-based LC/MS method was assessed over a concentration range of 0.1-10 ng/mL using the The recovery of the extraction method was calculated as 59%; however, this demonstrated good reproducibility as the %CV of the five recovery assessments was less than 5.3%. This indicated that although the recovery was low, the reproducibility of the recovery was high, and suggested that further optimisation could potentially increase the sensitivity of the method through increasing the elution efficiency of the peptide from the SPE cartridges.

| Quantification of INSL5 secretion in human and mouse colonic primary cultures
The primary colonic tissue cultures were incubated for 1 h with or without forskolin (10 μM), IBMX (10 μM) and glucose (10 mM), a cocktail well known to dramatically stimulate secretion of glucagon-like peptide-1 and peptideYY from L-cells, 4,16    from the supernatant and concentrating it significantly, such that the intact peptide could be detected and quantified using a high flow rate analysis approach. Whilst a high flow rate analysis is less sensitive than nano-LC/MS analyses, it is more robust and has significantly higher throughput. High flow analysis has been used previously to show how effective LC/MS/MS is as a quantitative tool for peptides and proteins, even generating comparable data against clinical analyser systems. 23 Unfortunately, the Phoenix murine INSL5 that was purchased for generating a quantitative assay also demonstrated a different sequence from the endogenous peptide; therefore, future LC/MS quantitative strategies need to be focused first on obtaining a modified reference standard. Further, and significant, methodology developments will be required for detecting and quantifying INSL5 in plasma, as we expect these to be in the pM (pg/mL) range, similar the intestinal hormones GLP-1 and PYY, which are co-secreted from L-cells. It is likely that in order to detect endogenous INSL5 in plasma, large plasma sample volumes could be required (not feasible in rodent models); however, nano-LC/MS/MS and targeted SRM detection could achieve the required sensitivity with limited volume samples. Furthermore, the investigation of a hybrid immuno-affinity LC/MS/MS analysis approach might be required, which was successfully employed to quantify circulating proglucagon products in human plasma. 13 However, here we demonstrated an approach to measure secreted and cellular INSL5 levels in mouse and human samples that should enable further research investigating the production and secretion of this peptide recently implicated to play a role in metabolic homeostasis.