ATP-mediated Events in Peritubular Cells Contribute to Sterile Testicular Inflammation

Peritubular myoid cells, which form the walls of seminiferous tubules in the testis, are functionally unexplored. While they transport sperm and contribute to the spermatogonial stem cell niche, specifically their emerging role in the immune surveillance of the testis and in male infertility remains to be studied. Recently, cytokine production and activation of Toll-like receptors (TLRs) were uncovered in cultured peritubular cells. We now show that human peritubular cells express purinergic receptors P2RX4 and P2RX7, which are functionally linked to TLRs, with P2RX4 being the prevalent ATP-gated ion channel. Subsequent ATP treatment of cultured peritubular cells resulted in up-regulated (pro-)inflammatory cytokine expression and secretion, while characteristic peritubular proteins, that is smooth muscle cell markers and extracellular matrix molecules, decreased. These findings indicate that extracellular ATP may act as danger molecule on peritubular cells, able to promote inflammatory responses in the testicular environment.

Male infertility is common and in a considerable number of cases the underlying causes are not known 1,2 . In infertile men, impairments of spermatogenesis are typically paralleled by alterations of testicular morphology. Common changes include fibrotic thickening of the tubular wall, and accumulation of macrophages and mast cells in both the testicular interstitial area and the tubular wall [3][4][5][6] . These alterations point to a form of sterile inflammation in the testes, specifically prevalent in the tubular wall, which is formed by peritubular cells and extracellular matrix (ECM).
Peritubular myoid cells are smooth muscle-like cells known for their contractile abilities that are of utmost importance for sperm transport 7,8 . Previous studies, including proteomic and secretomic analyses, revealed that these human testicular peritubular cells (HTPCs) secrete ECM components and act as paracrine signalling cells 9 . Intriguingly, they also secrete immunoregulatory factors 10 . Recently, Toll-like receptors (TLRs) as functional key regulators of innate immune responses were identified in HTPCs 11 . It became evident that ligands like Pam3CysSerLys4 (PAM) or lipopolysaccharide (LPS) are able to activate TLR2/4 on peritubular cells. In addition, TLR2/4 was also targeted by the small ECM molecule biglycan in the same way as previously found in macrophages 12 . Biglycan-induced TLR signalling triggered an immune response including pro-inflammatory cytokine production and secretion 13,14 .
In this context, simultaneous activation of TLR2/4 and the purinergic receptor isoforms P2RX4 and P2RX7 by biglycan has been discovered 15 . Both, P2RX4 and P2RX7, represent members of a family of ligand-gated ion channels that are activated by ATP at either relatively low (P2X4; EC 50~1 -10 µM) or substantially increased (P2X7; EC 50~1 00-300 µM) concentrations 16 . In the testis, potential origins of extracellular ATP are infiltrating immune cells like mast cells and macrophages, as well as Sertoli cells 17,18 . Both cell types reside in the immediate vicinity of peritubular cells 3,19,20 . Thus, we hypothesized that ATP may act as a danger molecule in the testes in the context of sterile inflammation and may promote inflammatory responses in HTPCs. We explored this possibility in a human-focused approach.
HTPCs are sensitive to purinergic stimulation in vitro. To address whether HTPCs could be part of a paracrine purinergic signalling network 22 , we performed whole-cell patch-clamp recordings from cultured cells  (1)(2)(3)(4) and in the human testis (+). Patient-derived HTPCs were additionally screened for the presence of smooth muscle cell markers ACTA2 and CNN1 and absence of the mast cell marker TPSAB1. Negative controls: non-reverse transcription control (−RT), non-template control (−). (b) Immunoblotting confirmed P2RX4 and P2RX7 expression in HTPCs from three different patients (a-c). Note that gel/blot images were cropped and full-length gels/blots are presented in Supplementary Fig. 5. (c) Basal P2RX4 (n = 8) and P2RX7 (n = 8/6) receptor mRNA levels at 6 h and 24 h varied between cells derived from individual patients, but also ACTA2 levels (n = 8) did not remain constant. Data are geometric mean with 95% confidence interval. (d) P2RX4 expression was detected in peritubular cells, germ cells and interstitial tissue, while P2RX7 expression was found in peritubular cells and vessels (not shown) solely. P2RX4 expression in the tubular wall overlapped with SMA and CNN1 expression and absence of tryptase staining in consecutive sections. Insets: Negative controls (pre-adsorption for P2RX4, omission of primary antibody for P2RX7); Bars = 20 µm.
SCIenTIFIC REPORTS | (2018) 8:1431 | DOI:10.1038/s41598-018-19624-3 ( Fig. 2a). HTPCs exhibited an average membrane capacitance (C mem ) of 27.5 ± 8.3 pF. When cells were exposed to ATP (10-1,000 µM) at a negative holding potential (V hold = −80 mV), we consistently recorded a fast-activating inward current (I ATP ) in 47.5% of all HTPCs, whereas no such current was recorded in a similarly large fraction of cells (52.5%; Fig. 2b i-ii ). Upon prolonged exposure (5 s), the current saturated and then monotonically declined in presence of the stimulus (Fig. 2b i ). Next, we asked whether ATP sensitivity is dose-dependent. At increasing extracellular ATP concentrations ([ATP] ex ) of 10, 100 and 1,000 µM, HTPCs exhibited dose-dependent I ATP amplitudes with apparent saturation at [ATP] ex ≥ 100 µM (Fig. 2c). In sharp contrast to previous observations in mouse spermatogonia 23 , the relatively small, desensitizing inward currents initiated by lower [ATP] ex (10-100 µM) did not change in amplitude, kinetics or desensitization rate upon exposure to high stimulus concentrations (1,000 µM). To examine the current-voltage relationship of I ATP , we measured I-V curves at a rate of 2 Hz before, during and after stimulation with saturating [ATP] ex (100 µM; Fig. 2d-f). Currents reversed at approximately 0 mV and exhibited substantial inward rectification (Fig. 2d-f). The above properties all indicate that P2RX7 does not serve as the predominant ATP receptor in HTPCs [24][25][26] . Rather, our data suggest that cultured HTPCs express one or more purinergic receptor isoform(s) characterized by relatively high ATP sensitivity.
HTPC purinergic signals are likely mediated by P2RX4 receptors. To increase throughput, we next opted to analyse ATP sensitivity using ratiometric Ca 2+ imaging of HTPC populations. For recording of cytosolic Ca 2+ concentrations [Ca 2+ ] c , HTPCs were loaded with fura-2/AM and briefly (3 s) exposed to increasing doses of ATP (0.1-1,000 µM; Fig. 3a). In many cells (33.5%), ATP stimulation triggered an immediate and robust [Ca 2+ ] c signal that decayed after stimulus cessation (Fig. 3b). It is likely that the size of this responsive population is an underestimate since many of the remaining HTPCs (66.5%) were precluded from analysis because of spontaneous activity (see Material and Methods). When both peak [Ca 2+ ] c signals (Fig. 3c) and response frequencies (Fig. 3d) were plotted as a function of stimulus concentration, ATP consistently induced dose-dependent responses with a shared activation threshold of ~3 µM. The linear dynamic range of the dose-response curves (Fig. 3c, 25,27,28 , these results also point to expression of purinoceptor isoform(s) that display high ATP sensitivity.
Next, we sought to identify the involved P2X receptor isoforms by pharmacological fingerprinting of ATP-induced [Ca 2+ ] c signals (Fig. 4). First, HTPCs were exposed repeatedly to saturating ATP concentrations (1,000 µM) in presence of standard (1 mM; S 1 ) versus reduced (165 nM; S 2 ) free extracellular [Ca 2+ ]. Notably, we never observed ATP-induced [Ca 2+ ] c elevations when extracellular [Ca 2+ ] was reduced (Fig. 4a,e). This finding shows that, first, Ca 2+ influx from the extracellular medium is necessary for ATP-mediated responses and that, second, P2Y receptor-dependent depletion of intracellular stores does not serve a primary function in HTPC purinergic signalling. Next, stimulations were carried out in presence of the selective P2RX7 receptor antagonist A438079 (10 µM), which blocks human P2RX7 at nanomolar concentrations 26 but, in HTPCs, failed to affect ATP-mediated signals (Fig. 4b,e).

ATP mediates changes of peritubular cell characteristics and fosters up-regulation of inflammatory molecules.
HTPCs are smooth muscle-like cells, which also secrete extracellular matrix, contribute to the spermatogonial stem cell niche and exhibit immunoregulatory functions 6 . To assess the impact of ATP on these characteristic features of HTPCs, three different strategies were applied: evaluation of changes in transcript levels, regulation of secreted cytokines in the supernatant and a whole cell proteomic approach. Cell viability during treatment was monitored via cell number and confluence determination in a live cell imaging system and additionally by cytotoxicity measurement via lactate dehydrogenase (LDH) release. Results revealed that cellular viability was not negatively influenced by ATP ( Supplementary Fig. 3).
ATP treatment resulted in a distinct decrease of purinoceptor mRNA expression after 6 h and 24 h (Fig. 5a). Characteristic smooth muscle cell markers (ACTA2, CNN1) were similarly decreased (Fig. 5b). ATP affected stem-cell niche regulatory factors (CXCL12, GDNF) only marginally, but in a time-dependent fashion (Fig. 5c). However, a panel of inflammation-associated genes showed moderately (IL6, IL33, CCL2) to substantially (IL1B, CCL7; Fig. 5d) increased expression. Yet, there was a different time-dependence in the mRNA elevation. While most transcript levels rose until 24 h, IL6 clearly reached a peak before 24 h. These changes suggest that ATP induces a switch from the smooth muscle cell-like phenotype and towards an immune-regulatory phenotype as it has been observed in vascular smooth muscle cells 31 .
Cytokine profiling of ATP-treated supernatants further revealed purinergic regulation of cytokine secretion in HTPCs (Fig. 6a,c). Firstly, we analysed those cytokines previously observed on transcript level. IL1B and IL33 signals were below detection limit and IL6 and CCL2 were only marginally up-regulated. In comparison, levels of the immunosuppressive cytokines, e.g. IL3, IL4 or IL10, were not affected (not shown). By contrast, CCL7 abundance appeared numerically increased in treated HTPCs. In addition, we identified another factor, CXCL5, with elevated expression after ATP treatment, yet this increase did not reach statistical significance (Fig. 6a). CXCL5 mRNA expression levels in ATP-treated samples were however statistically significantly augmented at 24 h (Fig. 6b). Among secreted proteins with reduced abundance after ATP treatment three previously unexpected candidates were determined: insulin-like growth factor-binding protein 3 (IGFBP3), osteopontin (SPP1) and thrombospondin 1 (THBS1), although only the first mentioned two factors were statistically significantly decreased (Fig. 6c). Here, concomitant detection of mRNA decrease reinforced our findings except for IGFBP3 (Fig. 6d).
The third method, mass spectrometry of whole cell pellets, identified 83 in abundance affected candidate proteins (p-value < 0.05; log2-fold change > |0.6|, Supplementary Table 1) in ATP-treated HTPCs (n = 6, 48 h), 38 were reduced in abundance and 45 were increased in abundance. All proteins that were quantified (3533 proteins, identification with two individual peptides, false discovery rate (FDR) <1%) are listed in Supplementary Table 1. Interaction annotation revealed clustering between proteins of decreased abundance (Fig. 6e, PPI enrichment p-value 2.61e −11 ), but not among those of increased abundance. For instance, six KEGG pathways were significantly enriched, the most prominent being ECM-receptor interaction (8 proteins, FDR 8.04e −10 ). Enriched GO terms were mostly related to collagen metabolism and ECM organization as well (Supplementary Table 1). To corroborate this finding, transcript abundances of several of the clustering proteins were compared in ATP-treated versus untreated HTPCs (Fig. 6f). Significant decreases in collagens COL1A1, COL1A2, COL3A1, COL4A2, COL6A2 and lysyl oxidase (LOX) mRNA after 6 h corresponded to the results of the proteomic study.
All three approaches yielded different, but yet consistent results. Elevated cytokine production and secretion in response to ATP reinforce the proposition of ATP as a danger molecule able to promote (pro-)inflammatory processes in HTPCs. Meanwhile smooth muscle-like cell markers and ECM-secretive properties of the HTPCs were negatively influenced.

Discussion
Extracellular ATP serves as a danger molecule in a variety of tissues 32 , but its role in the human testis in health and disease has been poorly examined. To explore its actions in the testicular environment of men we turned to HTPCs, to date the sole human testicular cell culture model available, which provides the possibility of mechanistic studies 6 . We applied complementary approaches and found that ATP can evoke distinct changes in HTPCs, which can foster inflammation. Therefore elevated levels of testicular ATP may contribute to male infertility. In the testis, extracellular ATP can originate from different sources, including Sertoli cells 33 and activated immune cells 18,34 . Strikingly increased numbers of activated mast cells have previously been reported in testes of infertile patients, specifically in the peritubular compartment 3,20 . We confirmed and correlated this finding with the expression of the purinergic receptors P2RX4 and P2RX7 in the tubular wall in vivo and in vitro. A recent study in the lung identified a link between increased extracellular ATP levels, P2RX7 expression and development of fibrosis 35 . Together, these findings may hint to a role of ATP in sterile inflammatory events associated with male infertility.
Notably, we observed inter-individual differences in P2X receptor expression. This variability may result from patient's variety and lifestyle, which cannot be controlled for, but also from the extent of impaired testicular function. These fluctuations in expression levels may explain variations in ATP-induced gene expression and cytokine release.
Electrophysiological characteristics of cultured HTPCs stemming from two different men pointed to P2RX4 as the prevalent purinergic receptor compared to P2RX7. Measurements of intracellular [Ca 2+ ] changes supported the hypothesis of a P2RX-mediated effect since ATP-induced Ca 2+ signals depended on the presence of extracellular Ca 2+ . Both a specific P2RX7 antagonists and a high affinity blocker for P2RX1-3 and P2RX5 failed to substantially inhibit ATP-dependent [Ca 2+ ] elevations, whereas micromolar concentrations of PPADS dose-dependently blocked such signals. Consistent with the electrophysiological properties I ATP in HTPCs, our findings suggest predominant functional expression of P2RX4. However, co-assembly and co-activation with P2RX7 cannot be excluded 36,37 and require future investigation.
To date, peritubular cells have primarily been characterized by the expression of smooth muscle cell markers 21,38 and loss of contractile markers has been associated with infertility 39 . Hence, ATP-mediated ACTA2 and CNN1 decreases in HTPCs could be first steps en route to infertility. By contrast, prior studies of HTPCs revealed their ability to secrete immunoregulatory factors 10 . Recently, functional Toll-like receptors were identified in HTPCs and a danger signal from the ECM, biglycan, was implicated in their activation, entailing secretion of the chemo-/cytokines CCL2 and IL6 11 . ATP was also able to induce CCL2 and IL6 production and secretion in HTPCs, similar to results from other cell types [42][43][44][45] . However, IL6 and CCL2 levels were only mildly elevated. In contrast, CCL7, a sister molecule of CCL2, exhibited a strong increase in response to ATP. A comparable result has previously been reported in murine mast cells 46 . IL1B is the prototypical pro-inflammatory cytokine and highly elevated by extracellular ATP in HTPCs and other cells e.g. via purinergic mechanisms [47][48][49] . Its family member IL33 is constitutively expressed by smooth muscle cells 50 and plays an important role during inflammation in diseases associated with tissue fibrosis 51 . CXCL5 is a novel player in ATP-mediated cytokine regulation, as it has mainly been known as a chemoattractant in neutrophil recruitment 52 . Yet, it has the ability to bind to the same receptor as CCL2 and CCL7 and may be jointly regulated 53,54 . Together, ATP evoked an increase in (pro-)inflammatory and immunoregulatory gene expression and corresponding protein secretion in HTPCs.
All identified factors reduced in abundance by ATP are important components of the ECM 55 . While collagens mainly form the ECM core structure, THBS1, SPP1 and IGFBP3 exhibit various interactive properties. THBS1 is able to bind to collagens and other ECM molecules, including biglycan, thereby modulating cell-matrix interactions 56 . SPP1 is expressed in numerous cell types including smooth muscle cells 57 and is involved as pro-inflammatory factor in macrophage recruiting and cytokine secretion 58 . Moreover, THBS1 and SPP1 have been described to directly bind to IGFBP3 (albeit with lower affinity than to IGFBP5, which was not included in our secretion array) 59 . Increased IGFBP3, apart from its role in regulation of soluble insulin-like growth factor availability, can also induce elevated collagen production in smooth muscle cells 60 . This could imply the possibility of a collagen down-regulation upon a decrease in IGFBP3.
We are not aware of comparable in vitro studies examining ATP action on ECM in other human smooth muscle cells, yet the result of ATP-induced ECM molecule decreases in HTPCs contrast in vivo findings from systemic studies of both the kidney and respiratory tract in mice 61,62 . It therefore remains to be determined whether and, if so, how the ATP/P2RX4 axis of peritubular cells may be related to testicular fibrotic ECM deposits in men suffering from infertility. Such ECM deposits could conceivably develop as a consequence of an overall inflammatory environment in vivo. Temporal aspects may also play a crucial role. Since in our study the mRNA decrease was observed mostly at 6 h, but was restored at 24 h, only a transient decline in ECM production may occur. Limited access to patient material prevented us from further investigation.
In summary, the experimental approaches delivered complementary results, which taken together provide mechanistic insights into the actions of ATP in the human testis. Specifically, elevated cytokine production and secretion in response to ATP reinforce the proposition of ATP acting via P2X receptors as a danger molecule able to promote (pro-)inflammatory processes in HTPCs.

RT-PCR and qPCR.
Electrophysiology. Cultured cells were transferred to the stage of an inverse microscope (DMI 4000B, Leica Microsystems, Wetzlar, Germany) equipped for phase contrast and infrared-optimized differential interference contrast (IR-DIC). Cells were observed using phase contrast objectives (HC PL FLUOTAR 10 × 0.3 NA and HCX PL FL L 63 × 0.70 NA CORR PH2; Leica Microsystems). Images were recorded using a cooled CCD-camera (DFC360FX; Leica Microsystems). Patch pipettes (5-7 MΩ) were pulled from borosilicate glass capillaries (1.50 mm OD/1.05 mm ID; Science Products) on a PC-10 micropipette puller (Narishige Instruments, Tokyo, Japan), fire-polished (MF-830 Microforge; Narishige Instruments) and filled with pipette solution (S 3 or S 4 depending on experimental design). An agar bridge (150 mM KCl) connected reference electrode and bath solution. An EPC-10 USB amplifier controlled by Patchmaster v2x90.1 software (HEKA Elektronik) was used for data acquisition. We monitored and compensated pipette and membrane capacitance (C mem ) as well as series resistance (R series = 16.0 ± 7.8 MΩ (mean ± SD)). Liquid junction potentials were calculated using JPCalcW software 70 and corrected online. Signals were low-pass filtered [analog 3-and 4-pole Bessel filters (−3 dB); adjusted to 1 / 4 − 1 / 5 of the sampling rate (10 kHz; depending on protocol)]. If not stated otherwise, holding potential (V hold ) was −40 mV. All data were recorded at RT. Voltage ramp protocols are described in the Results section.
Notably, cells that exhibited spontaneous recurrent Ca 2+ transients could not be unequivocally categorized and were therefore discarded. Thus, the percentage of ATP-sensitive HTPCs that we report here is likely to represent an underestimate.
Data analysis. Statistical analyses of qPCR and supernatant data were obtained using GraphPad Prism 6.0 Software (GraphPad Software Inc., San Diego, CA, USA). qPCR data were analysed via a one-sample t-tests of −ΔΔCq values, as were cytokine levels in the supernatant via a one-sample t-tests of not normalized values. Electrophysiology and Ca 2+ imaging data were obtained from independent experiments performed on at least two days with cells derived from two different patients. Individual numbers of cells/experiments (n) are denoted in the figure and/or captions. Data were analysed offline using FitMaster 2.9 (HEKA Elektronik), IGOR Pro 6.4 (WaveMetrics, Lake Oswego, OR, USA) and Excel 2013 (