*Article* **Triiodothyronine or Antioxidants Block the Inhibitory Effects of BDE-47 and BDE-49 on Axonal Growth in Rat Hippocampal Neuron-Glia Co-Cultures**

**Hao Chen † , Rhianna K. Carty † , Adrienne C. Bautista , Keri A. Hayakawa and Pamela J. Lein \***

> Department of Molecular Biosciences, University of California, Davis, CA 95616, USA; hachen@ionisph.com (H.C.); rhianna.k.carty@gmail.com (R.K.C.); abcashion@ucdavis.edu (A.C.B.); kahayakawa@ucdavis.edu (K.A.H.)

**\*** Correspondence: pjlein@ucdavis.edu; Tel.: +1-530-752-1970

† These authors contributed equally to this work.

**Abstract:** We previously demonstrated that polybrominated diphenyl ethers (PBDEs) inhibit the growth of axons in primary rat hippocampal neurons. Here, we test the hypothesis that PBDE effects on axonal morphogenesis are mediated by thyroid hormone and/or reactive oxygen species (ROS)-dependent mechanisms. Axonal growth and ROS were quantified in primary neuronal-glial cocultures dissociated from neonatal rat hippocampi exposed to nM concentrations of BDE-47 or BDE-49 in the absence or presence of triiodothyronine (T3; 3–30 nM), N-acetyl-cysteine (NAC; 100 µM), or α-tocopherol (100 µM). Co-exposure to T3 or either antioxidant prevented inhibition of axonal growth in hippocampal cultures exposed to BDE-47 or BDE-49. T3 supplementation in cultures not exposed to PBDEs did not alter axonal growth. T3 did, however, prevent PBDE-induced ROS generation and alterations in mitochondrial metabolism. Collectively, our data indicate that PBDEs inhibit axonal growth via ROS-dependent mechanisms, and that T3 protects axonal growth by inhibiting PBDEinduced ROS. These observations suggest that co-exposure to endocrine disruptors that decrease TH signaling in the brain may increase vulnerability to the adverse effects of developmental PBDE exposure on axonal morphogenesis.

**Keywords:** axonal growth; developmental neurotoxicity; neuronal morphogenesis; PBDE; reactive oxygen species; thyroid hormone

#### **1. Introduction**

The brominated flame retardants, polybrominated diphenyl ethers (PBDEs), are considered to be likely environmental risk factors for neurodevelopmental disorders [1–4]. Epidemiologic studies have identified a negative association between developmental exposure to PBDEs and executive function, motor behavior, and attention in infants and children [5–12]. These findings are of significant public health concern given the documented widespread human exposure to PBDEs with significantly higher body burdens in infants and toddlers relative to adults [13,14]. However, there remains significant uncertainty regarding the underlying mechanism(s) by which PBDEs interfere with neurodevelopment.

It has been hypothesized that PBDE developmental neurotoxicity reflects altered patterns of neuronal connectivity [12,15,16]. A critical determinant of the patterns of connections formed between neurons during development is axonal morphology. Interference with temporal and/or spatial aspects of axonal morphogenesis has been shown to cause functional deficits in experimental models [17–19]. Moreover, altered patterns of axonal growth are implicated in the pathogenesis of various neurodevelopmental disorders [20,21]. Recently, we demonstrated that BDE-47, a PBDE congener that is highly abundant in human tissues, and BDE-49, an understudied PBDE congener with levels

**Citation:** Chen, H.; Carty, R.K.; Bautista, A.C.; Hayakawa, K.A.; Lein, P.J. Triiodothyronine or Antioxidants Block the Inhibitory Effects of BDE-47 and BDE-49 on Axonal Growth in Rat Hippocampal Neuron-Glia Co-Cultures. *Toxics* **2022**, *10*, 92. https://doi.org/10.3390/ toxics10020092

Academic Editor: Remco H.S. Westerink

Received: 23 December 2021 Accepted: 2 February 2022 Published: 18 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

comparable to BDE-47 in gestational tissues of women living in southeast Michigan [22], inhibited axonal growth in primary hippocampal neuron-glia co-cultures, in part by delaying neuronal polarization [23].

BDE-47 and BDE-49 effects on axonal growth in primary hippocampal neurons were prevented by pharmacological blockade of ryanodine receptors (RyR) or siRNA knockdown of RyR, implicating RyR-dependent mechanisms in PBDE developmental neurotoxicity [23]. However, an unexpected finding from our previous studies was that the axon inhibitory effects of BDE-47 and BDE-49 exhibited comparable concentration-effect relationships despite significant differences in their potency at the RyR [24]. This observation raised the possibility that the RyR is not the primary molecular target but rather a downstream effector in the adverse outcome pathway (AOP) linking PBDEs to axonal growth inhibition. PBDEs have been shown to interfere with thyroid hormone (TH) signaling and to cause oxidative stress via increased levels of intracellular reactive oxygen species (ROS) [25,26], and both TH and ROS are reported to modulate RyR activity [27] and to influence axonal growth [28,29]. Therefore, in this study, we leveraged a primary rat hippocampal neuron-glia co-culture model to assess the relative contributions of TH and ROS-dependent mechanisms in mediating the axon inhibitory activity of BDE-49 and BDE-47. Our findings support the hypothesis that PBDEs inhibit axonal growth via ROS-dependent mechanisms, and that the TH, triiodothyronine (T3), protects against the effects of PBDEs on axonal growth by blocking PBDE-induced ROS.

#### **2. Materials and Methods**

#### *2.1. Materials*

Neat certified BDE-47 (2,2′ ,4,4′ -tetrabromodiphenyl ether, >99% pure) and BDE-49 (2,2′ ,4,5′ -tetrabromodiphenyl ether, >99% pure) were purchased from AccuStandard Inc. (New Haven, CT, USA), and verified for purity and composition by GC/MS by the UC Davis Superfund Research Program Analytical Core. Stock solutions of each BDE were made in dry dimethyl sulfoxide (DMSO, Sigma-Aldrich, St. Louis, MO, USA). 3,3′ ,5-Triiodo-L-thyronine (T3), N-acetyl-L-cysteine (NAC) and DL-α-tocopherol acetate were purchased from Sigma-Aldrich.

#### *2.2. Animals*

All procedures involving animals were approved by the University of California Davis Animal Care and Use Committee and conformed to the NIH Guide for the Care and Use of Laboratory Animals, and the ARRIVE guidelines [30]. Timed-pregnant Sprague Dawley rats were purchased from Charles River Laboratory (Hollister, CA, USA) and individually housed in clear plastic cages with corn cob bedding at 22 ± 2 ◦C under a 12 h dark–light cycle. Food and water were provided ad libitum.

#### *2.3. Cell Culture*

Primary neuron-glia co-cultures were prepared from hippocampi harvested from postnatal day (P) 0–1 male and female rat pups as previously described [31]. Briefly, rat pups were separated from the dam and anesthetized by placing them on a gauze pad on ice. Once pups ceased moving, they were euthanized by decapitation using sterile scissors. Hippocampi were harvested from the pup's head by sterile dissection and then dissociated using trypsin (1 mg/mL) and DNAse (0.3 mg/mL). Dissociated hippocampal cells were plated on poly-L-lysine (0.5 mg/ML, Sigma Aldrich) coated glass coverslips (BellCo, Vineland, NJ, USA) and maintained at 37 ◦C in NeuralQ Basal Medium supplemented with 2% (*v*/*v*) GS21 (MTI-GlobalStem, Gaithersburg, MD, USA) and GlutaMAX (ThermoScientific, Waltham, MA, USA). The concentration of T3 in the complete medium used to maintain cultures was ~2.6 nM [32,33]. For studies of axonal growth, neurons were plated at 27,000 cells/cm<sup>2</sup> ; for qPCR and Western blot experiments, neurons were plated at 105,000 cells/cm<sup>2</sup> . Cultures were exposed to varying concentrations of BDE-47 or BDE-49 diluted in culture medium from 1000× stocks; vehicle control cultures were exposed to

DMSO (1:1000 dilution). A subset of cultures was co-exposed to T3, NAC, or α-tocopherol diluted 1:1000 directly into cell cultures from 1000× stocks in sterile distilled water.

#### *2.4. Quantification of Axonal Outgrowth*

Cultures were exposed to BDE-47, BDE-49, or vehicle (1:1000 DMSO) for 48 h beginning 3 h post-plating, and then fixed with 4% (*w*/*v*) paraformaldehyde (Sigma Aldrich) in 0.2 M phosphate buffer. To visualize axons, hippocampal cultures were immunostained with antibody specific for tau-1 (1:1000, Millipore, Billerica, MA, USA, RRID AB\_94855). Our previous studies [23] demonstrated that exposure to BDE-47 or BDE-49 did not alter the expression of tau, as determined by Western blotting. Axonal lengths of tau-1 immunopositive neurons were manually quantified by an individual blinded to experimental condition using ImageJ software with the NeuronJ plugin [34]. As previously defined [35], in any given neuron, the axon was identified as the neurite whose length was >2.5× the cell body diameter and exceeded that of the other minor processes of the same neuron. Only non-overlapping neurons were quantified as proximity to other neurons can affect neuronal morphology.

#### *2.5. Quantitative Polymerase Chain Reaction (qPCR)*

Total RNA was isolated from cell cultures using TRIzol Reagent (ThermoScientific) per the manufacturer's instructions, and cDNA was synthesized using the SuperScriptTMViloTM MasterMix containing SuperScriptTM III Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA). Samples were mixed with Power SYBR Green MasterMix and forward and reverse primers (see Supplemental Table S1 for primer sequences and amplification efficiencies) and then loaded into a MicroAmp 384 Reaction Plate (ThermoScientific). qPCR plates were run on a 7900HT System by the Real-Time PCR Research and Diagnostics Core Facility at UC Davis. qPCR primers and probes were ordered from Integrated DNA Technologies (Coralville, IA, USA) using PrimeTime® Predesigned qPCR Assays. Transcript levels were normalized to the average of the reference genes Ppia and Hprt1 and expression ratios were calculated by Pfaffl method [36] using REST 2009 software (Qiagen, Valencia, CA, USA).

#### *2.6. ROS Measurements*

Rat hippocampal neurons cultures were exposed to BDE-47, BDE-49, or vehicle (1:1000 DMSO) in the absence or presence of T3, NAC, or α-tocopherol 3 h post-plating. Global ROS production was measured 1 h following exposures using ROS-Glo assay (Promega, Madison, WI, USA) according to manufacturer's protocol, which specified using H2O<sup>2</sup> as a positive technical control. Luminescence was recording using an H1 hybrid microplate reader (BioTek Instruments, Winooski, VT, USA).

#### *2.7. Mitochondrial Metabolism Kinetics*

Primary rat hippocampal neuron cultures were plated in 96-well plates at 27,000 cells/cm<sup>2</sup> for 48 h. Cells were then exposed to BDE-47, BDE49, or vehicle (1:1000) in the absence or presence of T3 in combination with a mitochondrial substrate library, MitoPlate-S (Biolog, Inc., Hayward, CA, USA). Mitochondrial substrate metabolism was characterized according to the manufacturer's protocol. Kinetics was recorded on the H1 hybrid microplate reader at a wavelength of 590 nm.

#### *2.8. Statistics*

All data are presented as mean ± SE unless otherwise indicated. Graphs were created in GraphPad Prism 8.3.0. Statistical analyses were performed with GraphPad Prism using one-way ANOVA with post hoc Tukey's or Dunnett's or post hoc Kruskal–Wallis with Dunn's as appropriate for the normality of the data as measured by Shapiro–Wilk. qPCR data were analyzed using SDS 2.4 (ThermoScientific) and REST 2009 software (Qiagen, Valencia, CA, USA) with statistical analyses performed using REST 2009 pairwise reallocation randomization test. Significant differences between single and co-exposures or positive

controls and vehicle were determined using Student's *t*-test. Statistical significance was defined as *p* < 0.05.

#### **3. Results**

#### *3.1. T3 Blocked the Axon Inhibitory Effects of BDE-47 and BDE-49*

We previously demonstrated that exposure to either BDE-47 or BDE-49 at concentrations ranging from 200 pM to 2 µM inhibited axonal growth in primary rat hippocampal neurons [23]. To address the question of whether these PBDE congeners modulated axonal growth via effects on TH signaling, we first tested whether the axon inhibitory activity of PBDEs could be blocked by supplementation of the culture medium with T3. Axon lengths were quantified on day in vitro (DIV) 2 after a 48 h exposure to BDE-47 or BDE-49 at 2 or 200 nM in the absence or presence of exogenous T3 at 3 or 30 nM. Consistent with our previous findings, BDE-47 or BDE-49 did not alter the number of axons extended by an individual neuron, but these PBDEs did significantly reduce axonal length relative to vehicle controls (Figure 1A,B). Addition of exogenous T3 at 3 or 30 nM, which raised T3 concentrations in the culture medium to ~5.6 and 32.6 nM, respectively, prevented the inhibition of axonal growth by BDE-47 or BDE-49, as indicated by the fact that axon lengths of neurons exposed to PBDEs in culture medium supplemented with T3 were not significantly different from those of vehicle controls (Figure 1A,B).

**Figure 1.** T3 supplementation prevented BDE-47 and BDE-49 inhibition of axonal growth in primary hippocampal neurons. Primary neuron-glia co-cultures dissociated from the hippocampi of P0-1 rats were exposed to vehicle (DMSO diluted 1:1000) or varying concentrations of BDE-47 or BDE-49 in the absence or presence of T3 beginning 3 h after plating. After 48 h exposure, cultures were fixed and immunostained for the axon-selective cytoskeletal protein tau-1. (**A**) Representative photomicrographs

of DIV 2 hippocampal neurons exposed to vehicle, BDE 47 at 2 nM ± exogenous T3 at 3 nM. Scale bar = 25 µm. (**B**) Quantification of axon length in tau-1 immunopositive neurons. Data presented as the mean ± SE (*n* = 70–90 neurons from three independent dissections). \*\*\* Significantly different from vehicle at *p* < 0.001; # significantly different from the corresponding BDE treatment in the absence of T3 at *p* < 0.05 as determined by one-way ANOVA followed by Tukey's post hoc test. (**C**) Fold-change in transcript levels of *Klf9* (as a % of vehicle control). Data are presented as the mean ± SE of *Klf9* expression normalized to the average of the reference genes *Ppia* and *Hprt1*. \* Significantly different from vehicle at *p* < 0.05 as determined by REST 2009 pairwise randomization test.

T3 is a component of many neuronal cell culture medias [34,35], and the medium used in these studies contained T3 at ~2.6 nM [34,35]. Thus, our observation that T3 supplementation protected against PBDE inhibition of axonal growth raised the possibility that PBDEs inhibited axonal growth by interfering with TH signaling. As one test of this possibility, we determined whether PBDEs interfered with TH-mediated gene expression. The gene Kruppel-like factor 9 (*Klf9*), previously known as Basic transcription element-binding protein (*Bteb*), has been shown to be a sensitive TH-responsive gene in the developing brain [36,37]. Analyses of *Klf9* transcripts in 2 DIV hippocampal cell cultures confirmed that *Klf9* expression is significantly upregulated in cultures exposed to exogenous T3 at 3 nM for 48 h (Figure 1C). In contrast, exposure to BDE-47 or BDE-49 at 200 nM for 48 h had no significant effect on *Klf9* transcript levels relative to vehicle control cultures and did not inhibit the upregulation of *Klf9* by T3 (Figure 1C).

To determine whether the protective effect of T3 on PBDE inhibition of axonal growth was mediated via direct effects of T3 on axonal growth, we quantified the effect of supplementing the culture medium with T3 on axonal growth in cultures not exposed to PBDEs. As seen in representative photomicrographs (Figure 2A), supplementation with T3 at either 3 or 30 nM had no obvious effect on axonal morphology in terms of the number, length, or branching of axons in DIV 2 hippocampal neurons. Quantification of axon length confirmed that 48 h exposure to medium supplemented with T3 did not significantly alter axon length relative to that observed in vehicle control cultures (Figure 2B).

**Figure 2.** T3 did not influence axonal growth. Primary neuron-glia co-cultures dissociated from the hippocampi of P0-1 rat hippocampi were exposed to vehicle (DMSO diluted 1:1000) or T3 and/or

α

α

α

α

BDE-47 or BDE-49 beginning 3 h after plating. After 48 h exposure, cultures were fixed and immunostained for tau-1. Representative photomicrographs (**A**) and quantification of axon length (**B**) in tau-1 immunopositive neurons at DIV 2. Data are presented as the mean ± SE (*n* = 30–40 neurons per group from one dissection; results repeated in 3 independent dissections). There were no significant differences between neurons exposed to vehicle vs. T3 as determined by one-way ANOVA (*p* < 0.05). Scale bar = 25 µm.

#### *3.2. Antioxidants Blocked PBDE Inhibition of Axonal Growth*

Previous reports have demonstrated that PBDEs increase levels of ROS in cultured neurons [38–40] and that PBDE-induced ROS can be blocked by mechanistically diverse antioxidants, specifically the NADPH oxidase inhibitor, NAC, or the ROS scavenger, α-tocopherol [41,42]. To evaluate a role for ROS in the axon inhibitory effects of PBDEs, we thus determined whether co-exposure to NAC or α-tocopherol blocked the inhibition of axonal growth by BDE-47 or BDE-49. No significant changes in axon length were observed with antioxidant treatment alone (Supplemental material Figure S1). As shown in representative photomicrographs (Figure 3A) and confirmed by quantitative morphometric analyses of axons (Figure 3B), axon lengths of hippocampal neurons exposed to BDE-47 or BDE-49 at 200 nM in the presence of 100 µM NAC or 100 µM α-tocopherol were not significantly different from those of vehicle control neurons. Cultures co-exposed to PBDEs and antioxidants were significantly longer than axon lengths of hippocampal neurons exposed to the corresponding BDE alone.

**Figure 3.** *Cont*.

α

α **Figure 3.** Antioxidants prevented BDE-47 and BDE-49 inhibition of axonal growth and production of ROS. Primary neuron-glia co-cultures dissociated from the hippocampi of P0-1 rat pups were exposed to vehicle, BDE-47 or BDE-49 in the absence or presence of N-acetyl cysteine (NAC) or α-tocopherol. After 48 h exposure, cultures were fixed and immunostained for tau-1. (**A**) Representative photomicrographs of DIV 2 hippocampal neurons from different experimental groups. Scale bar = 25 µm. (**B**) Quantification of axon length in tau-1 immunopositive cells (*n* = 70–90 neurons from three independent dissections). Quantification of ROS levels following exposure to vehicle, BDE-47 or BDE-49 alone (**C**) or in the presence of an antioxidant (**D**) (*n* = three independent dissections). H2O<sup>2</sup> was included as a positive technical control for the ROS-Glo assay per the manufacturer's instructions. Data presented as the mean ± SE. \* Significantly different from vehicle at \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001; # significantly different from PBDE treatment alone at # *p* < 0.05, ## *p* < 0.01, as determined by one-way ANOVA followed by Tukey's post hoc test; †significantly different from individual PBDE treatment at *p* < 0.05 as determined by Student's *t*-test.

To determine whether nM concentrations of BDE-47 or BDE-49 that inhibit axonal growth increased intracellular ROS, ROS were measured in cultures acutely exposed to BDE-47 or BDE-49. Both BDE-47 and BDE-49-exposed cultures had higher amounts of ROS compared to vehicle control cultures (Figure 3C). We next evaluated whether antioxidants blocked the inhibitory effects of PBDEs on axonal growth by providing protection against ROS generation (Figure 3D). In the presence of NAC or α-tocopherol, PBDEs did not produce significant amounts of ROS compared to vehicle. However, ROS production was substantially reduced compared to BDE-47 or BDE-49 alone.

It is posited that ROS generation largely originates from mitochondrial damage [43]. BDE-47 can disrupt the mitochondrial membrane potential [44], while both BDE-47 [45,46] and BDE-49 [47] can decrease mitochondrial bioenergetics. Thus, we next sought to determine whether acute exposure to nM concentrations of BDE-47 or BDE-49 altered mitochondrial metabolism. Compared to vehicle control cultures, mitochondrial metabolism was significantly impacted in cultures acutely exposed to either BDE-47 or BDE-49 at 200 nM (Figure 4B).

#### *3.3. T3 Blocked PBDE Axon Inhibition by Blocking PBDE-Induced ROS*

To determine whether T3 conferred protection against the axon inhibitory effects of PBDEs via upregulation of endogenous antioxidant molecules, we quantified the effects of T3, BDE-47, and BDE-49, alone and in combination, on the production of ROS (Figure 4A). In contrast to cultures exposed to PBDEs in the absence of T3, in cultures co-exposed for 1 h to one of these PBDEs and T3 exhibited no significant change in ROS levels relative to vehicle controls. Moreover, ROS levels were significantly reduced in cultures co-exposed to PBDEs and T3 relative to cultures exposed to PBDEs in the absence of T3. We then explored whether T3 protected against disrupted mitochondrial bioenergetics (Figure 4B). Following acute exposure to either BDE-47 or BDE-49 in combination with T3, there were no marked alterations in mitochondrial substrate metabolism relative to vehicle. In addition, any effects observed with individual PBDE exposure were eliminated in cultures co-exposed to PBDEs and T3.

tt ff **Figure 4.** T3 normalized ROS levels and mitochondrial substrate metabolism in cultures exposed to BDE-47 or BDE-49. Hippocampal neuron-glia co-cultures were exposed to vehicle, T3, BDE-47 and/or BDE-49 for 1 h on DIV 2. (**A**) Quantification of ROS production following co-exposure to T3 and PBDEs. (**B**) Mitochondrial substrate metabolism kinetics immediately following PBDE exposure alone and in the presence of T3. Data presented as the mean ± SE (*n* = three independent dissections). \* Significantly different from vehicle at \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001; # significantly different from T3 at # *p* < 0.05, ## *p* < 0.01, ### *p* < 0.001, #### *p* < 0.0001 as determined by one-way ANOVA followed by Dunnett's post hoc test; †significantly different from individual PBDE treatment at † *p* < 0.05, ††† *p* < 0.001 as determined by Student's *t*-test.

#### **4. Discussion**

The findings from this study extend our previous report that BDE-47 and BDE-49 inhibited axonal growth in primary rat hippocampal neurons [23] by demonstrating that the axon inhibitory activity of these PBDE congeners is mediated by increased levels of intracellular ROS. The evidence in support of this conclusion includes: (1) BDE-47 and BDE-49 increased ROS in primary rat hippocampal neurons at nM concentrations that also inhibited axonal growth; and (2) co-exposure to either the NADPH oxidase inhibitor, NAC, or the ROS scavenger, α-tocopherol, blocked the axon inhibitory effects of BDE-47 and BDE-49. Additionally, we observed that supplementation of the culture medium with exogenous T3 blocked the inhibition of axonal growth in PBDE-exposed neuronal cultures, coincident with mitigation of PBDE effects on intracellular ROS and metabolic substrate production from the mitochondria. These findings suggest a role for T3 in maintaining intracellular redox homeostasis in response to pro-oxidants, which if true, represents a novel mechanism by which thyroid hormone disruption contributes to adverse neurodevelopmental outcomes.

Our observations are consistent with previous reports that PBDEs upregulated biomarkers of oxidative stress in the brain of adult and developing rodent models [48,49] and increased intracellular ROS levels in cultured neural cells [38,39,50,51]. This earlier work demonstrated that µM concentrations of PBDEs increased intracellular ROS in cultured neurons to levels that triggered apoptosis [24,52,53]. Here, we found that exposure of primary hippocampal neuron-glia co-cultures to BDE-47 or BDE-49 at nM concentrations also increased intracellular ROS, but this was associated with inhibited axonal growth. Our data extend reports in the literature indicating that physiologic levels of ROS regulate axonal specification and axonal growth in primary hippocampal neurons, and modulation of ROS synthesis in axonal growth cones cause cytoskeletal rearrangements that alter axonal morphogenesis [54]. Collectively, these observations suggest a model in which nM PBDE concentrations increase ROS locally in the axonal growth cone to modulate signaling pathways that regulate axonal growth [55,56], whereas µM PBDE concentrations increase intracellular ROS globally to trigger cell death. Confirmation of this model will require the adaptation of sensitive technologies to detect localized changes in ROS in subcellular domains of neurons [57] exposed to PBDEs at concentrations that inhibit axonal growth.

PBDEs can interfere with thyroid hormone signaling and thyroid hormone disruption is widely posited to contribute to the developmental neurotoxicity of these environmental contaminants [25,26]. PBDEs have been shown to suppress dendritic growth in Purkinje cells by disrupting TH receptor-mediated transcription [58], and we observed that coexposure to T3 blocked inhibition of axonal growth by BDE-47 or BDE-49. However, several lines of evidence argue against the hypothesis that PBDEs inhibit axonal growth in hippocampal neurons via direct interference with TH signaling. First, in hippocampal cultures not exposed to PBDEs, T3 supplementation of the culture medium did not promote axonal growth. Second, exposure of hippocampal cultures to BDE-47 or BDE-49 did not alter expression of *Klf9*, a gene known to be highly sensitive to upregulation by TH in the developing brain [59]. Nor did BDE-47 or BDE-49 significantly block T3-induced *Klf9* expression. These findings are in agreement with previous studies [58] in which qPCR analyses detected no significant changes in transcript levels of TH-responsive genes, including TRα1 or TRβ, in primary rat Purkinje cells exposed to PBDEs. Moreover, since the affinity of T3 to the thyroid hormone receptor (THR) is approximately 0.1 nM, the observation that T3 present in the medium without addition of extra T3 is not sufficient to prevent the axon inhibitory effects of PBDEs suggests that the neuroprotective effect of exogenous T3 in this model is mediated by THR-independent mechanisms. NH-3, a pharmacological THR modulator with mixed agonist/antagonistic activity [60], may be useful for addressing this question, but given experimental evidence that the concentration– response relationship for antagonistic vs. agonist effects of NH-3 vary across models, its effectiveness in mechanistic studies of the axon inhibitory activity of PBDEs will require identification of a concentration that antagonizes THR in this model system [60,61].

Our data suggest that T3 supplementation prevented PBDE inhibition of axonal growth by mitigating PBDE-induced ROS. Specifically, we observed that T3 supplementation ameliorated PBDE-induced ROS generation. A key question is how. TH has been reported to upregulate expression of endogenous antioxidant molecules [62,63]. However, preliminary qPCR analyses failed to detect significant upregulation of several endogenous antioxidants in primary hippocampal neuron-glia co-cultures exposed to BDE-47 or BDE-49 in the presence of T3 (Supplemental Table S2). This observation does not rule out the possibility that T3 upregulated expression of cellular antioxidants other than those we assessed and/or that T3 increased the activity of enzymatic antioxidants. In addition, *Klf9* upregulation by 5 or 10 nM T3 supplementation has previously been shown to protect the axons of primary cortical murine neurons from hypoxic injury [64]. Whether *Klf9* or other T3-regulated targets are directly involved in mitigating PBDE axon inhibition remains to be investigated. However, it is now clear that TH can also signal via non-transcriptional

mechanisms [65–67], including direct influence on mitochondrial respiration [68]. Consistent with this literature, our data support a model in which T3 protects mitochondrial metabolism against PBDE-mediated disruption of mitochondrial bioenergetics.

The observation that T3 prevented PBDE-induced changes in mitochondrial substrate utilization at concentrations that also blocked PBDE inhibition of axon growth, yet had no effect on basal axonogenesis, suggested that PBDEs increased ROS as a consequence of altered mitochondrial metabolism. In support of this proposed mechanism, at concentrations that increased intracellular ROS levels, BDE-47 and BDE-49 increased utilization of metabolic substrates (α-keto-isocaproic acid, α-keto-butyric acid, Ala-Gln, D-glucose-6-PO4) used to produce NADH, and disruption of NADH production has been linked to increased ROS generation [69]. However, the mechanism(s) by which PBDEs interfere with mitochondrial metabolism remain to be elucidated.

Findings from our previous studies suggested RyR was a downstream effector in PBDE-induced axon growth inhibition [23]. Given the redox-sensitive nature of RyR [70,71] and spatial relationship with mitochondria [72], a potential indirect mechanism presents itself wherein mitochondrial ROS production alters RyR gating and, consequently, calcium signaling to interfere with axon growth. This model is supported by experimental evidence demonstrating that disruption of mitochondrial function affects calcium homeostasis, which in turn delays polarization of developing neurons and inhibits axonal growth [73]. As we previously reported [23], PBDE inhibition of axonal growth is due in part to delayed neuronal polarization. The role of RyR as a downstream key event rather than the molecular initiating event in PBDE developmental neurotoxicity may explain the differential response of dendrites vs. axons to non-dioxin-like polychlorinated biphenyls (PCBs) vs. PBDEs. Specifically, in primary rat hippocampal and cortical neuron-glia co-cultures, non-dioxinlike PCBs were observed to promote dendritic growth, but have no effect on axonal growth, and the dendrite promoting activity was mediated by RyR sensitization [74,75]. In contrast, PBDEs were observed to inhibit axonal growth but have no effect on dendritic growth in the same neuronal cell culture model [23].

In summary, our study provides novel insight into the interplay between ROS, TH, and axonal growth in PBDE developmental neurotoxicity. Whether PBDE interference with axonal growth contributes to adverse neurodevelopmental outcomes in vivo is still to be determined; however, clinical [20,21] and experimental evidence [17–19] demonstrate that altered spatiotemporal patterns of axonal growth during brain development can cause functional deficits. Susceptibility to this neurotoxic activity of PBDEs may be enhanced in populations with heritable mutations that alter mitochondrial and redox signaling, which are themselves associated with increased risk of neurodevelopmental disorders [76,77]. The finding that T3 protects against axon growth inhibition by BDE-47 and BDE-49 in vitro suggests that PBDE-mediated TH dysregulation [2,78] also has the potential to amplify PBDE effects on axonal growth in vivo. Further studies into gene × environment interactions associated with these mechanisms may lead to a better understanding of populations with increased vulnerability to PBDE developmental neurotoxicity.

**Supplementary Materials:** The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/toxics10020092/s1, Figure S1: The antioxidants NAC and α-tocopherol do not alter axonal outgrowth relative to vehicle control cultures. Primary neuron-glia co-cultures dissociated from the hippocampi of P0-1 rat pups were exposed to vehicle, N-acetyl cysteine (NAC) or α-tocopherol. After a 48 h exposure, cultures were fixed and immunostained for tau-1. Axon length was quantified in tau-1 immunopositive cells (*n* = 70–90 neurons from three independent dissections). Data presented as the mean ± SE. No significant differences between groups was detected using oneway ANOVA (*p* < 0.05); Table S1: Primer sequences and amplification efficiencies; Table S2: Average fold changes relative to vehicle of levels of transcripts encoding cellular antioxidants.

**Author Contributions:** Conceptualization, H.C., R.K.C. and P.J.L.; data curation, H.C., R.K.C. and A.C.B.; formal analysis, H.C. and R.K.C.; funding acquisition, P.J.L.; investigation, H.C., R.K.C., A.C.B. and K.A.H.; methodology, H.C., R.K.C. and K.A.H.; project administration, P.J.L.; supervision, P.J.L.; visualization, H.C. and R.K.C.; writing—original draft, H.C. and R.K.C.; writing—review and editing, H.C., R.K.C., A.C.B., K.A.H. and P.J.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the National Institute of Environmental Health Sciences, grant numbers R01 ES014901, P30 ES023513, P42 ES04699, and T32 ES007059, and the National Institute of Child Health and Development, grant number P50 HD103526. The contents of this work do not necessarily represent the official views of the NIEHS or NICHD, and these institutes do not endorse the purchase of any commercial products or services mentioned in the publication.

**Institutional Review Board Statement:** The animal study protocol was approved by the Institutional Animal Care and Use Committee of the University of California, Davis (protocol #18813 and #18853, approved 18 April 2013).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data will be made available upon reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


**Nelson T. Peterson <sup>1</sup> and Chad M. Vezina 1,2, \***


**Abstract:** Lower urinary tract dysfunction (LUTD) is nearly ubiquitous in men of advancing age and exerts substantial physical, mental, social, and financial costs to society. While a large body of research is focused on the molecular, genetic, and epigenetic underpinnings of the disease, little research has been dedicated to the influence of environmental chemicals on disease initiation, progression, or severity. Despite a few recent studies indicating a potential developmental origin of male LUTD linked to chemical exposures in the womb, it remains a grossly understudied endpoint in toxicology research. Therefore, we direct this review to toxicologists who are considering male LUTD as a new aspect of chemical toxicity studies. We focus on the LUTD disease process in men, as well as in the male mouse as a leading research model. To introduce the disease process, we describe the physiology of the male lower urinary tract and the cellular composition of lower urinary tract tissues. We discuss known and suspected mechanisms of male LUTD and examples of environmental chemicals acting through these mechanisms to contribute to LUTD. We also describe mouse models of LUTD and endpoints to diagnose, characterize, and quantify LUTD in men and mice.

**Keywords:** lower urinary tract dysfunction; lower urinary tract symptoms; BPH; prostate

#### **1. Introduction**

LUTD is a deviation from normal urinary voiding. While LUTD occurs in males and females, disease mechanisms differ between sexes. The prostate plays a considerable role in male LUTD, the focus of this review. For such a pervasive disease, male LUTD has suffered from a surprising lack of research attention. Part of the problem is the disease's complexity, driven by a constellation of underlying factors across multiple organs that are incompletely understood. Another problem is that the historical research record for LUTD is muddled by vast and inconsistent nomenclature used to describe the disease, decentralizing the resource of primary peer-reviewed literature. Several vocabulary terms are used to describe histological, anatomical, physiological, and clinical pathologies in the lower urinary tract. The following terms are sometimes conflated or interchanged with LUTD, and often used inappropriately: benign prostatic hyperplasia (BPH), benign prostatic enlargement (BPE), bladder outlet obstruction (BOO), partial bladder outlet obstruction (pBOO), lower urinary tract symptoms (LUTS), and others. These terms are defined in Table 1.

Male LUTD can be confirmed by specialized urodynamic studies at the urology clinic (diagnostic and experimental approaches used to identify LUTD mechanisms in mice and humans are described in Table 2). However, male LUTD is most often identified in the primary care clinic based on patient reported symptoms. LUTS can include but are not limited to weak stream, incomplete bladder emptying and more frequent voiding, especially at night. Male LUTD frequently begins in the fifth decade of life or later and is a progressive disease that can result in a loss of bladder function, bladder and kidney stones, acute urinary retention, and renal injury/failure [1–7]. LUTD disrupts sleep and

**Citation:** Peterson, N.T.; Vezina, C.M. Male Lower Urinary Tract Dysfunction: An Underrepresented Endpoint in Toxicology Research. *Toxics* **2022**, *10*, 89. https://doi.org/ 10.3390/toxics10020089

Academic Editor: Soisungwan Satarug

Received: 4 January 2022 Accepted: 11 February 2022 Published: 16 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

has been linked to depression, decreased workplace productivity, and a reduced quality of life [8–12]. If not successfully managed, LUTD can be fatal.

**Table 1.** Definitions of terms used to describe anatomical and physiological disorders of the male lower urinary tract.



**Table 2.** Strengths and limitations of methods to evaluate lower urinary tract dysfunction in men and male mice.


**Table 2.** *Cont.*

LUTD is extremely common. A 2008 study estimated that 1.9 billion people, representing 45% of the population, are affected by LUTD [9]. The economic burden of LUTD is staggering. The disease affects more than half of men over 50 years of age in the Western world, resulting in \$4 billion for the pharmacological treatment and \$2 billion for the surgical treatment of LUTD and associated prostatic problems [13–15]. The most common therapies for male LUTD are directed to block alpha adrenoreceptor function (alpha blockers) and dihydrotestosterone synthesis (steroid 5 alpha reductase inhibitors), factors which contribute to prostatic smooth muscle contraction and prostatic enlargement, respectively. Unfortunately, these therapies are incompletely effective. Their magnitude of effect is marginal, not all patients respond, and existing therapies are only moderately protective against disease progression [16–18]. It is becoming clear that male LUTD derives from many different mechanisms, not all of which are addressed by current therapies. Factors responsible for severe drug-refractory disease are not understood. Recent studies reveal potential roles for environmental chemical exposures, during the fetal period when the lower urinary tract is developing [19–21] and during other stages, in driving LUTD susceptibility and progression, opening an entirely new line of toxicology research towards understanding environmental factors that contribute to LUTD processes.

This review is intended as a resource for toxicologists and other discipline specialists who are considering entry into the urologic disease research space and wishing to examine LUTD as a toxicology research endpoint. We describe the anatomy, cellular composition, and physiology of male lower urinary tract organs including the bladder, urethra, and prostate. We describe known and emerging disease mechanisms. We also highlight the limited examples of how environmental chemicals influence male LUTD and list opportunities for future research.

#### **2. Overview of Male Lower Urinary Tract Anatomy and Physiology**

Several benign diseases of the lower urinary tract are accompanied by a change in distribution, type, or state of cells that comprise lower urinary tract tissues [22–24]. Therefore, we describe the cellular anatomy of the male lower urinary tract to give toxicologists an appreciation of the normal cellular organization and changes which occur in response to chemical insults and disease The male lower urinary tract consists of the bladder, prostate, and urethra (Figure 1). Urine flows from the kidney to the bladder via the ureter and passes through the prostatic urethra and prostate before continuing through the penile urethra and exiting the body as voided urine (Figure 1).

**Figure 1.** General anatomy of the male urinary tract and effects of chemical insults on the male lower urinary tract. (**A**) A general depiction of the male lower urinary tract. (**B**) Known effects of environmental chemicals on the lower urinary tract of either the man or male mouse.

#### *2.1. The Bladder*

The bladder's primary functions are to store and expel urine. The bladder wall consists of three tissue layers: a specialized epithelium known as the urothelium, the lamina propria, and the bladder smooth muscle (detrusor) [25,26].

The mature urothelium is comprised of basal, intermediate, and superficial cells [27]. Bladder epithelial cell differentiation begins early in fetal development (weeks 7–8 in humans), and the trajectory of urothelial cell differentiation during development and regeneration is susceptible to epigenetic modification [28] revealing a potential mechanism of toxicity for epigenetic modifying chemicals. The mature urothelium must achieve

three unique functions. The first is to maintain distensibility to accommodate bladder filling and emptying. Bladder volume increases significantly during the storage cycle, a process which would normally challenge the integrity of an epithelial lining [29]. Bladder distensibility is achieved by urothelial cell junction rearrangements and cell sliding during bladder filling [29].

The second role of the urothelium is to protect sub-urothelial tissue from toxins, microorganisms, and urine solutes [27,30]. Barrier function is facilitated by secreted uroplakins [31]. Uroplakins are transmembrane proteins which assemble into a crystalline structure and are interrupted by hinge regions to allow bladder distension [32]. Uroplakins assemble to form uroplaques, rigid bio-membrane structures which cover 90% of bladder lumen [32]. Uroplaques are integral to the integrity, flexibility, and solubility of the urothelium [33]. The control of urothelial cell division is integral to maintaining functional uroplaques and restoring them after bladder damage. Although the urothelial cell turnover is normally slow with a labeling index of 1% in mice, the urothelium is reconstituted quickly after injury through the progenitor activities of basal and intermediate cells [32–36]. The epithelium of the urothelium can be completely repaired in 4 weeks in guinea pigs and 6 weeks in men [36]. Some mechanisms by which the bladder restores barrier function are surprising. For example, we found that under certain circumstances when widespread urothelial cell death depletes the bladder of its own progenitors, it can recruit non-resident, non-bladder (Wolffian duct) epithelial progenitors, drive their differentiation into uroplakin secreting superficial cells and restore barrier function [37]. Barrier function is crucial because sub-epithelial bladder cells are severely compromised by urine exposure. The experimental use of cyclophosphamide, an antineoplastic used therapeutically for Hodgkin's lymphoma, multiple myeloma, and other cancers, has widened the understanding of barrier function and consequences of barrier function loss. Cyclophosphamide is bio-transformed into acrolein, which accumulates in the urine and drives urothelial cell death, resulting in hemorrhagic cystitis and changes in physiology [38–40]. Environmental chemicals with urothelial cell toxicity are expected to drive bladder inflammation and dysfunction like that of cyclophosphamide.

The third role of the urothelium is that of a sensor. In combination with nerve terminals within the bladder, the urothelium detects and responds to mechanical and chemical stimuli to alter detrusor contractility and moderate bladder afferent nerve activity [41,42]. Factors released by urothelial cells include acetylcholine, adenosine triphosphate (ATP), nerve growth factor, nitric oxide (NO), prostaglandins, and others [41,43].

The lamina propria contains a fibroelastic connective tissue with intervening afferent and efferent nerve fibers, a vast vascular network and dispersed fibroblasts, a loose smooth muscle layer (the muscularis mucosa), and myofibroblasts [26,44]. The elastic fibers within the lamina propria allow the bladder to recover its original shape after voiding [45].

The detrusor is the major smooth muscle component of the bladder [46]. The detrusor is organized as a circular muscle inner layer sandwiched between longitudinal muscle outer layers [46]. Muscle bundles are surrounded by collagen [46–48]. Detrusor contraction is predominantly controlled by cholinergic neurons [49,50], but can also be induced by purinergic neurons and relatively rare sympathetic neurons [49,50].

The normal voiding cycle is divided into filling and voiding phases [51]. Urine expands the bladder during the filling phase, while bladder pressure remains lower than urethral pressure [50,51]. There is still uncertainty about how the bladder relays the perception of fullness to the brain. One possibility is that mechanoreceptors and mechanosensitive ion channels within the bladder transmit information about fullness to afferent neurons [52–56]. There is also evidence that urothelial cells, stretched during bladder filling, release ATP to activate purinergic receptors on bladder afferents and relay bladder fullness to the brain [57–59]. Another possibility is that the perception of fullness is not driven by a slow increase in bladder pressure (intravesicular pressure), but rather by an increasing rate of spontaneous transient contractions, also called micromotions, which exist throughout the filling phase. Micromotions drive the major portion of afferent outflow to the central nervous system

during bladder filling, acting in part through a mechanism involving calcium-activated potassium (SK type) channels [60].

In 1925, F.J.F. Barrington identified a brain stem region which controls micturition, including sensation of bladder fullness and the contractions leading to urination [61]. Studies using retrograde and anterograde neuronal labeling pinpointed the location of this micturition center in the pontine tegmentum [62–67]. This site of micturition control is referred to as Barrington's nucleus, the pontine micturition center, and the M-region [62]. Afferent and efferent urinary voiding pathways are integrated in Barrington's nucleus. During the storage phase, glutamatergic neurons in the periaqueductal gray and hypothalamus relay information about bladder fullness and bladder volume threshold for voiding to Barrington's nucleus [68,69]. During the voiding phase, corticotropin releasing hormone-positive and estrogen receptor 1-positive neurons within Barrington's nucleus activate efferent pathways to drive detrusor contraction [62,70,71]. Additional neurons in Barrington's nucleus send inhibitory signals to the external urethral sphincter, driving its relaxation and allowing urine to flow unimpeded from the bladder into the urethra [71–73]. Though there is widespread evidence that environmental contaminants can disrupt connectivity, complexity, arborization, and signaling of neurons within the peripheral and central nervous system, whether environmental chemicals impact bladder ascending and descending neural pathways is rarely examined [74–80].

There is limited evidence that environmental chemical exposures can disrupt bladder neural circuitry as it is established during the fetal and neonatal periods, raising concerns about a developmental basis of bladder health and disease. A recent study tested the impact of exposure to a polychlorinated biphenyl (PCB) mixture on bladder structure and function [19]. The PCB mixture used in this study mimics the most encountered congeners in women who are at risk for having a child with a neurodevelopmental disorder [81, 82]. PCBs were delivered orally to nulliparous female mice (75% C57BL/6J/25% SVJ129) starting two weeks before mating, through pregnancy and lactation, and continuing in offspring before their bladders were analyzed at postnatal days 28–31. The PCB mixture increased densities of sub-urothelial beta-3 tubulin (general neural fiber marker) fibers and calcitonin gene-related peptide positive (peptidergic fiber marker) fibers in male mice but not female mice, and these changes were accompanied by an increase in male bladder volume [19], suggesting they were sufficient to drive a change in bladder function.

#### *2.2. The Urethra*

The human male urethra is divided into two parts, consisting of five segments: the anterior urethra (fossa, penile, and bulbar segments) and the posterior urethra (membranous and prostatic segments) [83]. The rodent male urethra is divided into two parts—penile and pelvic [84]. The human and rodent urethra are populated by epithelial cells, smooth and striated muscle cells, blood vessels, and sensory and motor neurons [85]. While the cellular components of the anterior/penile urethra have not been extensively characterized, single cell ribonucleic acid (RNA) sequencing approaches have been used to determine the cellular components of the prostatic urethra [24,86,87]. Urethral epithelium consists of club cells, hillock cells, basal epithelial cells, and neuroendocrine cells [86]. Urethral club and hillock cells were recently identified, but their functional characterization is incomplete and represents a future research opportunity. Lung club cells, which are transcriptionally like those in the urethra, act as progenitors and mediate anti-inflammatory and antioxidant processes [88–90]. Lung hillock cells, which are transcriptionally like those in the urethra, serve as progenitors, and participate in barrier function and immunomodulation [91,92].

#### *2.3. The Urethral Sphincter*

The urethral sphincter serves as a valve to regulate urine flow between the bladder and urethra [93]. During the bladder storage phase, urethral pressure exceeds bladder pressure to maintain continence [50]. During the voiding phase, the urethral sphincter falls open, urethral pressure decreases while bladder pressure increases, the urethra distends and

urine flows through the prostatic urethra and penile urethra to become voided urine [50]. The urethral sphincter is divided into two parts: the external sphincter and the internal urethral sphincter [93,94]. The external sphincter consists of striated muscle circumscribing the urethra and is under voluntary control [93,95]. The internal urethral sphincter is indistinct from the rest of the lower urinary tract smooth muscle (bladder smooth muscle is continuous with urethral and prostatic smooth muscle), but is physiologically defined by its autonomic regulation, connected via a reflex arc to the bladder [95,96]. Urethral smooth muscle is organized as a thin longitudinal superficial layer, a dense circular layer, and a thin longitudinal deep layer [94].

#### *2.4. The Prostate*

The prostate synthesizes a portion of the ejaculate [97]. Prostatic smooth muscle contracts during ejaculation to propel prostatic fluid into the urethra [98]. The prostatic urethra also distends to accommodate urine during voiding. Benign prostatic disease changes the prostate's histology and cellular composition and can prevent prostatic urethral distention during voiding, causing BOO, a common etiology for LUTD (defined in Table 1).

The human prostate is a spherical gland encapsulated by a fibromuscular sheath known as the prostatic capsule [24,96,99]. The base of the prostate is adjacent to the bladder and the prostatic urethra courses through its center [100]. The prostatic ductal network is like that of a branched tree: the main ducts drain directly into the urethra and divide into primary, secondary, and tertiary branches as they extend towards acini concentrated in the gland's periphery [101]. The human prostate is organized into zones, differing in cellular composition and responsiveness to disease, and includes the transition zone (most susceptible to histological BPH, defined in Table 1) [24,102], the central zone and the peripheral zone (most susceptible to prostate cancer) [100,102]. The rodent prostate, often used as a disease model for humans, is anatomically distinct from the human prostate in that it is not spherical, but instead divided into four bilaterally symmetrical lobes: the anterior, dorsal, lateral, and ventral prostate [102]. While spontaneous cancer is not observed in the mouse prostate, a variety of genetically engineered mouse models are susceptible to prostate cancer and disease incidence differs by lobe [103]. The mouse prostate gland develops BPH spontaneously with age, but lesions are diffuse, like those that contribute to clinical disease in the dog, and unlike nodular BPH in the humans [104,105]. The rodent prostate ductal network is organized as a branched tree, like that of the human prostate, but ducts are surrounded by a looser stroma than in human prostate and the rodent gland is encapsulated in a thin adventitia instead of the thick capsule that surrounds the human prostate.

Human prostatic epithelium is made up of luminal, basal cells; neuroendocrine, club and hillock cells are also present, but are rare in prostate compared to urethral epithelium [22,86,106]. Human prostate stroma consists of three smooth muscle cell types (peri-prostatic, vascular smooth muscle and pericyte), two fibroblast cell types (peri-epithelial and interstitial), leukocytes, endothelial cells, and sensory and autonomic nerve fibers [86]. Mouse prostate stroma contains three fibroblast cell subtypes distributed in distinct proximal–distal and lobe-specific patterns and smooth muscle [24,106]. The transcriptomes of mouse prostatic and urethral fibroblasts are like human interstitial fibroblasts [24]. However, mouse urethral and ductal fibroblasts evoke Wingless relatedintegration site (Wnt) and Transforming growth factor beta (TGFβ) signaling pathways that are less abundant in human prostate fibroblasts [24]. Human peri-epithelial fibroblasts instead express Wnt inhibitors that could buffer Wnt ligands produced by other stromal or epithelial cells [24]. Human prostatic fibroblasts are organized in layers that center around epithelial structures, while mouse prostatic fibroblasts are not layered and differ by lobe [24]. Human and mouse prostate fibroblasts are most abundant in the proximal regions of prostatic ducts and least abundant in acini in the distal regions [24,107].

The recent observation, derived from single cell RNA-sequencing data, that human and mouse prostate cellular landscapes are similar, is also supported by previous microarray data [108]. Similarly, mouse prostate organogenesis is like that of the human prostate [109]. These data support the use of mice as a relevant model species for studying cellular and molecular mechanisms of benign prostatic disease. The key to understanding the differences in prostate architecture and benign prostate hyperplasia manifestation between these species may lie in the function of the specialized prostate epithelial and stromal cells of these species [109].

Prostate disease can be detected by changes in the spatial distribution and frequency of prostate cells [24]. Prostate cell immunophenotyping has proven difficult, as disease processes frequently lead to changes in cell state and cell type that cannot be easily distinguished by simple immunohistochemical staining protocols. New and validated RNA-Sequencing approaches, as well as cell sorting protocols deriving from them, have recently been described [86] and will be essential for elucidating prostate cell functions in future studies.

#### **3. LUTD Mechanisms**

#### *3.1. Benign Prostatic Diseases*

A variety of benign prostatic conditions contribute to male LUTD, many of which are believed to cause LUTD by driving BOO (defined in Table 1). The impacts of BOO extend beyond the prostate and into the bladder. A prolonged intravesicular pressure increase and bladder contraction against resistance reprograms the bladder in a process known as bladder compensation: the detrusor becomes thicker [110], it undergoes functional changes in ion channel physiology [111] and efferent signaling is reprogrammed [112]. If BOO is not effectively addressed, the bladder decompensates, much like a heart undergoing hypertrophic cardiomyopathy: the detrusor thins, is replaced by fibrotic tissue, and becomes incapable of mounting an effective contraction to fully evacuate urine from the bladder. There is evidence in rabbits that bladder decompensation is at least partially reversed by relief of bladder outlet obstruction [113]. Recovery from BOO likely depends on the severity of bladder decompensation at the time of surgery [113–115]. Thus, BOO must be effectively addressed before it permanently impairs bladder function.

BPH is a leading cause of LUTD in men of advancing age. Human BPH is defined by prostate histology, specifically the presence of stromal, epithelial, or mixed nodules in the central and transition zones (Table 1) [22,116–120]. Small hyperplastic nodules can form as early as the 3rd decade of life and increase in frequency and volume with advancing age [121]. BPH mechanisms are not fully understood, but it has been hypothesized that BPH arises from a reawakening of embryonic signaling pathways [121] or disrupted homeostatic regulation of cell growth and death programs [116–120].

Aging-related changes in circulating testosterone and 17-beta-estradiol concentrations are another mechanism linked to male LUTD. Serum and prostate tissue concentrations of testosterone and 17-beta-estradiol change with age in men [122,123] and the changes are associated temporally and mechanistically with male LUTD [124–126]. Pharmacological alterations in testosterone and 17-beta-estradiol are a proven cause of LUTD in non-human male primates, canines, rats, and mice [124,127–133]. In mice, slow-release implants of testosterone and estradiol drive an increase in voiding frequency, a reduction in voided volume, an increase in collagen deposition, and a change in velocity of urine flow through the prostatic urethra [124]. The mechanism by which changes in circulating testosterone and 17-beta-estradiol drive voiding dysfunction are not clear but may include direct actions on the bladder [134,135], changes in prostatic desmin and smooth muscle actin content or function [136–139].

The fact that LUTD arises from natural changes in circulating sex hormone concentration raises questions about impacts of endocrine disrupting chemicals on male voiding function, and this area of toxicology research is in its infancy. For example, subcutaneous implants of the estrogenic chemical bisphenol A (BPA, 25 mg), combined with testosterone (2.5 mg) and given to C57BL/6N adult (6–8 weeks old) male mice, increase bladder mass and volume, increase voiding frequency, and reduce the volume of voided urine, suggestive

of BOO [140]. BPA may act more broadly in the lower urinary tract, affecting the bladder as well as the prostate. Delivery of BPA (0.05–0.5 mg/kg/day) to Pietrain × Duroc mixedbreed juvenile female pigs increases the number and thickness of vasoactive intestinal polypeptide (VIP) expressing neurons in the bladder wall [141], raising questions about the influence of BPA, and the larger class of environmental estrogens to which it belongs, on detrusor recovery after contraction.

Prostate inflammation, also called prostatitis (defined in Table 1), is extremely common and has been closely associated with LUTD. Approximately 50% of prostate biopsy, surgical or autopsy specimens harbor evidence of histological inflammation, most typically characterized as chronic (lymphocytic) inflammation [142]. The incidence of prostate histological inflammation is even higher (75%) in men with LUTD [143]. The presence of prostate inflammation in a biopsy specimen correlates with risk of symptomatic progression, urinary retention, and need for surgery [142,144–146]. A significant proportion of men with histologically defined prostate inflammation will develop urinary dysfunction [147]. Two placebo-controlled drug trials, Reduction by Dutasteride of Prostate Cancer Events (REDUCE) and Medical Therapy of Prostatic Symptoms (MTOPS), correlate histological prostate inflammation in human male prostate with increased prostate volume [144]. MTOPS study outcomes reveal that men with histological inflammation are more likely to progress to advanced LUTD, including acute urinary retention [144]. A separate study found that men with prostatitis were 2.4 times more likely to develop BPH and the presence of histological prostate inflammation in baseline biopsies was associated with 70% increased odds of requiring later treatment for LUTD [146]. Despite clear evidence that some environmental chemicals can drive inflammation and modulate autoimmunity, there is little information about environmental impacts on prostate inflammation and this represents a future opportunity that can be examined using immunohistochemical and physiological methods in Table 2.

There is a distinction between histological and clinical prostatitis: histological prostatitis is identified in histological tissue sections, while clinical prostatitis is diagnosed by physical examination, urinalysis, imaging, cystoscopy, or patient questionnaire (for example, The National Institute of Health Chronic Prostatitis Symptom Index (NIH-CPSI)) [148]. Clinical prostatitis accounts for a significant proportion of outpatient visits [149]. Clinical prostatitis includes acute and chronic bacterial prostatitis, nonbacterial prostatitis, and asymptomatic prostatitis [148].

Prostate fibrosis is a recently identified mechanism of male LUTD. Fibrosis is an abnormal, detrimental version of the wound-healing process and is characterized by collagen deposition and tissue stiffening [150]. Macoska et al. [151] were the first to report fibrosis in the human prostate and link collagen accumulation to tissue stiffness and LUTS severity. Subsequent reports linked prostate fibrosis to histological inflammation, LUTS, and resistance to a combination therapy of alpha blockers and 5 alpha reductase inhibitors [150,152]. Prostatic fibrosis is an evolutionarily conserved LUTD process, supported by the fact that collagens also accumulate within the prostates of aging intact dogs and mice [104,153]. Though triggers for prostate fibrosis are not fully known, and whether environmental contaminants drive prostate fibrosis has not been studied, prostatic fibrosis results from prostate inflammation secondary to *E. coli* infection or obesity in mice [154,155].

Prostatic smooth muscle dysfunction is the target of the most prescribed drug class for male LUTD, the alpha blockers, and can be studied experimentally using calcium flux assays and isometric contractility assays described in Table 2. A study by Baumgarten et al. [156] was the first to identify noradrenergic axons in the human prostate, a surprising discovery considering that autonomic outflow to the bladder is mediated instead by cholinergic axons. Receptor binding studies and isometric contractility assays showed that noradrenergic receptors in prostatic smooth muscle mediate prostate tissue contractility [157,158]. The outcomes of these studies ushered the hypothesis that prostatic smooth muscle hyperactivity impairs urine flow through the prostatic urethra to cause BOO in some men. While this hypothesis was the basis for developing alpha blockers for male LUTD, little research

has been directed at identifying mechanisms of prostatic smooth muscle dysfunction, most notably dysfunction mediated by environmental chemicals. This area remains ripe for scientific exploration. Prostatic smooth muscle contraction is controlled by autonomic neurons and aging is one factor that may contribute to changes in prostatic innervation [124,159]. There is emerging evidence that environmental chemicals can also change prostatic innervation to cause prostatic smooth muscle dysfunction, specifically by acting during the fetal and neonatal periods when prostate autonomic innervation is established. For example, we recently showed in C57BL/6J mice that gestational exposure to the widespread environmental contaminant TCDD (a single 1 µg/kg oral maternal dose on the 13th day of gestation) increases noradrenergic fiber density (nerve terminals) in the prostate of male mouse fetuses without changing the density of cholinergic or peptidergic fibers [160]. TCDD-induced prostatic noradrenergic hyperinnervation persists into adulthood and is coupled to hyperactivity of prostatic smooth muscle and abnormal urinary function in mice, including increased urinary frequency [160]. These findings are important because they support the concept that prostate neuroanatomical development is malleable, at least in mice, and that intrauterine chemical exposures can permanently reprogram prostate neuromuscular function to cause male LUTD in adulthood. In contrast, exposure to TCDD and other aryl hydrocarbon receptor agonists during adulthood appear to protect against BPH in men [161,162]. Differing consequences of aryl hydrocarbon receptor activation in the fetal period, versus adulthood, highlight the need to control for age in studies that examine potential impacts of environmental chemicals on urinary function and LUTD.

#### *3.2. Bladder Mechanisms of Male LUTD*

A variety of bladder conditions can lead to urinary dysfunction. This section describes the most common causes of male LUTD.

Overactive bladder is characterized by involuntary detrusor contraction. Consistent changes in animal models of overactive bladder include patchy denervation of the bladder, enlarged sensory neurons, hypertrophic dorsal root ganglia, and an enhanced spinal micturition reflex [163]. Overactive bladder is often characterized by sensory dysfunction [163]. There is a role for muscarinic M2 receptors in the severity of urinary urgency [163]. Some individuals with overactive bladder have a thicker bladder wall, suggesting overactive bladder may derive from BOO in some men [163,164].

The etiology of overactive bladder is multifactorial, deriving from three major mechanisms: myogenic factors, urotheliogenic factors, and neurogenic factors [41]. Myogenic factors contributing to overactive bladder include spontaneous detrusor contractions in response to bladder distension, ischemia, and changes in smooth muscle properties over time [41]. Neurogenic factors may include abnormal sensory processes, abnormal afferent excitability, or in some cases, damage, or abnormalities in central processing [41]. Dimethylaminopropionitrile, used in the manufacture of polyurethane, is an inhalation hazard that acts through a neurogenic mechanism to cause overactive bladder [165]. Methyl mercury also causes overactive bladder through what appears to be a neurogenic mechanism [166,167]. Damage to the urothelium can also cause overactive bladder, as rupture of urothelial cells releases factors that can drive detrusor contractility and micturition [41]. Biphenyl, used as a resin, a heat transfer medium, and an anti-fungal, is an example of an environmental chemical that causes urothelial cell damage and death [168].

Underactive bladder, also known as detrusor underactivity, is defined by detrusor contraction of inadequate strength, and results in prolonged or incomplete bladder emptying [169]. Patients with underactive bladder have a diminished sense of bladder fullness and are unable to mount forceful bladder contractions [170]. Underactive bladder can occur after episodic overactive bladder, reminiscent of bladder decompensation after BOO. In fact, there is a documented relationship between LUTD, underactive bladder, and fibrosis of the bladder [171]. The interstitial cells of Cajal, a specialized cell population with smooth muscle pacemaking activity, have been implicated in underactive bladder. The frequency of interstitial cells of Cajal is reduced in mice with underactive bladder and is associated with

reduced frequency and amplitude of detrusor contraction [172]. Rats driven by bladder outlet obstruction to develop underactive bladder are deficient in stem cell factor, a ligand for the receptor C-kit which controls proliferation and function of interstitial cells of Cajal, and an increase in stem cell factor restores detrusor contractility [172].

#### *3.3. Urethral Mechanisms of Male LUTD*

Detrusor sphincter dyssynergia is characterized by simultaneous contraction of the detrusor and urinary sphincter, thereby impairing urine outflow from the bladder [173]. Detrusor sphincter dyssynergia manifests in three distinct phenotypes: (Type 1) increased sphincter activity during detrusor contraction which then ceases, resulting in delayed urination, (Type 2) intermittent clonic contractions during voiding, resulting in intermittent stream, (Type 3) continuous sphincter activity during detrusor contraction, resulting in impaired voiding [174]. Detrusor sphincter dyssynergia is common in men with spinal cord injuries or multiple sclerosis and has the capability to drive bladder decompensation, elevate pressure in the ureter and pelvis, and cause hydronephrosis, renal scarring and terminal kidney failure [173,174].

Neurological disease commonly manifests in bladder dysfunction [175]. Autonomic nervous system lesions (stroke, tumor, traumatic spinal cord injury, myelopathies due to cervico-arthrosis spina bifida), disseminated lesions (Parkinson's disease, brain trauma, multiple sclerosis, meningo-encephalitis,) and peripheral neuropathies (diabetes mellitius) have all been identified as mechanisms of bladder dysfunction [175] and act in part by disrupting coordination between the detrusor, urinary sphincter, and central nervous system [173,174]. While there are many examples of environmental chemicals causing neuropathies, the consequences on lower urinary tract function are rarely examined.

#### *3.4. The Relationship between LUTD and Comorbidities*

Recent studies connect LUTD to other diseases. For example, people with cardiovascular disease, diabetes and obstructive sleep apnea are at increased risk of developing LUTD [176–179]. A common thread linking these diseases is a change in hemodynamics connected to ischemic injury [180–182], a factor that independently drives LUTD in mice [174]. Environmental chemical exposures have been linked to cardiovascular disease and diabetes [183–186], and this is another mechanism by which they may drive LUTD.

#### **4. Mouse Research Models of Male LUTD**

Here we describe animal models used to study various etiologies of LUTD. While it is important to realize that results from animal models are not always transferable to humans, it is also crucial to highlight that animal models are used in preclinical trails to test the safety and efficacy of drugs and are an invaluable tool to use in toxicological studies.

#### *4.1. Benign Prostatic Hyperplasia*

A variety of genetically engineered mouse models have been used to drive expression of growth factors or mitogenic hormones in the prostate. Androgen responsive promoter sequences, androgen-induced cre recombinase or viral promoters are used to target genetic modifications to mouse prostate tissue and overexpress fibroblast growth factor 2 or fibroblast growth factor 3 to drive epithelial BPH [187–189], overexpress prolactin [190–193] or interleukin 1 alpha [194] to drive epithelial and stromal BPH and prostate inflammation, delete serine/threonine kinase 11 to promote stromal BPH in the periurethral region [195], or genetically modify other sequences. Expression of an activated form of P110 alpha, the catalytic subunit of PI3K, in mouse prostate epithelium also drives epithelial BPH in mice but accompanied with a stark fibrotic response in prostatic stroma [196]. Many of these genetically engineered mouse models were created before contemporary methods were optimized for mouse urinary physiology phenotyping. The historical goal was to use genetically engineered mice to identify molecular mediators of BPH and test efficacy of drugs and dietary substances for relieving BPH in preclinical model species. While genetically engineered mouse models are useful for understanding homeostatic mechanisms of prostate cell proliferation, it is becoming clear that BPH is not always linked to LUTD in men [197], and it remains important to characterize urinary physiology in these mice as a more relevant endpoint for male LUTD.

#### *4.2. Mouse Models of Prostate Inflammation*

Histological inflammation of the human prostate is extremely common: in one study, it was detected in nearly 80% of prostate biopsy specimens from 60+ year old men and was strongly associated with urinary voiding symptoms [198]. Prostate infection by ascending microbes is one potential mechanism of prostate inflammation and supported by the frequent encounter of bacteria in human prostate tissue specimens [199–201]. One strategy for driving prostate inflammation in mice involves urethral catheterization and delivery of uropathogenic *E. coli*. A variety of isolates have been used (*E. coli* UTI89, 4017, 1677 and CP-1), ranging from those collected as clinical urine isolates from women with bladder infections, to others collected from men with pelvic pain [202,203]. The pattern of inflammation (acute vs. chronic) depends in part on mouse strain used [204] and method of *E. coli* delivery (single vs. multiple inoculations, catheter size, instillation volume and bacterial load). It is essential to control these variables carefully when considering experimental design, and mice instilled with sterile saline (sham operated mice) are an essential component of experimental design because urethral catheterization can itself induce trauma, urethritis, and changes in urinary voiding physiology [144]. Prostatic *E. coli* infection is linked to prostate fibrosis and changes in voiding patterns in mice, but voiding patterns differ between *E. coli* strains and methods of infection and can include high volume, low frequency voiding [144] or low volume, high frequency voiding [107,205,206].

Many men with histological prostatitis present with a pattern of prostatic infiltrate consistent with prostate autoimmunity [207–209], an observation co-opted for the design of mouse models. The prostate ovalbumin expressing transgenic-3 mouse expresses ovalbumin under the control of the androgen-responsive probasin promoter [210,211]. Autologous splenocytes are activated in vitro and transplanted to drive T-cell mediated prostate autoimmunity and inflammation [207]. While the pattern of inflammation and mechanisms of cell proliferation have been carefully studied in this mouse, the urinary physiology phenotype is not well characterized. The experimental autoimmune prostatitis mouse model involves repeated intradermal injections of rat prostate homogenate into mice to drive a T-cell based autoimmune reaction that has been used to examine mechanisms of male LUTD and chronic pelvic pain [212,213].

A non-bacterial mouse model of prostate inflammation was created based on observations that IL-1 beta abundance increases after intraprostatic injection of noxious agents or uropathogenic *E. coli* infection [214–217] and that Prostatic IL-1 beta abundance is elevated in humans with histological BPH and correlates with LUTS and chronic pelvic pain [218–222]. This mouse model utilizes the Tet-On system which induces expression of a gene in the presence of doxycycline and is tunable, with stronger transgene expression with doxycycline dose [223]. A double transgene of Hoxb13-rTA transgene and a TetO-IL1 beta responder is used to drive IL-1 beta in prostatic epithelial cells [223,224]. The urinary metabolomic proteomic signatures of this mouse have been described, but the urinary physiology phenotype remains to be determined [224,225].

A recent mouse model of prostate inflammation is based off observations that prostate secretory proteins are leaked into prostate stroma of some men with LUTD and accompanied with patchy loss of the adherens junction protein e-cadherin, suggesting a loss of prostate barrier function [226]. Genetic depletion of e-cadherin in mouse prostate epithelium increases prostate mass and cell proliferation, thickens prostate stroma, and increases voiding frequency while reducing voided urine volume, and increases spontaneous bladder contractions [227].

#### *4.3. Mouse Models of Partial Bladder Outlet Obstruction (pBOO)*

Surgical approaches were first used to model pBOO in male mice. One approach involves a retropubic incision to apply and cinch a suture or metal ring around the bladder neck or pelvic urethra to drive bladder compensation and overactive bladder, and later bladder decompensation, detrusor underactivity, fibrosis, and loss of muscle mass [228–231]. The mouse model has been essential for recognizing new druggable pathways for restoring function to the decompensated bladder [232].

Treatment with exogenous androgens combined with estrogens is a non-surgical method to drive BOO in mice. Mice are given slow-release implants of androgen (testosterone or dihydrotestosterone) in combination with slow-release implants of estrogen (17beta-estradiol or diethylstilbesterol). The combination of androgen plus estrogen is necessary for prostate gland maintenance, as estrogens delivered to male mice in the absence of androgens disrupts hypothalamic/pituitary/gonadal signaling and cause prostate gland atrophy [128]. Genetically engineered mice that overexpress aromatase are also used to recapitulate the endocrine environment of advancing age [129,233]. Male mice treated with androgen and estrogen develop progressive LUTD, with evidence of disease processes (increased bladder weight as evidence of hypertrophy/compensation for BOO) occurring as early as two weeks after treatment [124]. Sustained exposure to exogenous androgens and estrogens elicits a variety of changes to the male lower urinary tract of multiple species, including prostatic hypertrophy and inflammation, urethral narrowing and abnormal urethral muscle tone, urinary dysfunction with progressive onset, bladder overactivity and eventual decompensation [124,127–131,133,234–238]. Estrogen receptor activation is a key driver of urinary dysfunction, as exogenous estradiol given to male mice drives urinary retention in the absence of exogenous testosterone [239] and estrogen receptor 1 is required for urinary retention and voiding dysfunction from exogenous testosterone and 17beta-estradiol in mice [237]. An important consideration when exogenous androgen and estrogen are used to drive male LUTD, especially when incorporating genetic changes to identify mechanisms, is that hormone responsiveness, disease onset, progression and severity are influenced by genetic background and mouse strain [234]. The delivery system of exogenous androgens and estrogens should be considered if using hormones to drive LUTD for a toxicology study. Compressed pellets of androgens and estrogens can be crushed when animals are restrained for chemical exposure (injection) [21]. Silastic capsule preparations of androgens and estrogens are more durable [240].

#### *4.4. Mouse Models of Overactive Bladder (OAB)*

OAB can be induced by ischemic injury [24,93,241]. A balloon catheter is passed through the iliac artery and inflated, then withdrawn to cause endothelial damage [241]. This injury is combined with a cholesterol enriched diet to cause bladder arterial occlusions and chronic bladder ischemia [206]. This model results in increased voiding frequency but decreased voided volume, and more frequent non-voiding bladder contractions in rats [241].

OAB can also be induced by the introduction of noxious stimuli (acetic acid, hydrochloric acid, and others) into the bladder [242,243]. Chemical induced OAB decreases the inter-voiding interval of anesthetized mice, reduces bladder capacity, and sensitizes afferent nerves [243–246].

#### *4.5. Mouse Models of Detrusor Sphincter Dyssynergia (DSD)*

The most common approach to evoke detrusor sphincter dyssynergia is to induce spinal cord injury under anesthesia [247]. Urine must be manually expressed at least three times per day until the micturition reflexes recover (10–14 days), then once per day [247]. Cystometry profiling of injured mice reveals increased activity of the external urethral sphincter coupled with increased urethral pressure and voiding pressure, increased frequency and magnitude of non-voiding contractions, and increased bladder capacity [247,248].

#### **5. Conclusions**

Lower urinary physiology is extremely complex, shaped by contributions from the urethra, prostate, bladder, ascending and descending neural pathways, and the brain. Despite an extremely high prevalence of male LUTD and devastating impacts on society, LUTD mechanisms and factors that influence LUTD severity are poorly understood. Environmental contributions to LUTD remain almost completely unexamined. We provided this overview of male lower urinary tract anatomy, physiology, and cell biology, described known disease mechanisms, and highlighted knowledge gaps that require additional research to direct new attention from toxicologists and environmental health specialists to this widespread disease. We detailed examples of environmental chemicals that perturb urinary tract function and described mouse models of LUTD with the intention that public health specialists, epidemiologists and toxicologists will consider LUTD research in toxicity assessments. Future risk mitigation strategies will likely be critical to reducing the burden and severity of LUTD in aging adults.

**Author Contributions:** Conceptualization, N.T.P. and C.M.V. writing—original draft preparation, N.T.P. and C.M.V.; writing—review and editing, N.T.P. and C.M.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by National Institutes of Health Grants R01 ES001332 and U54 DK104310, T32 ES007015, and the UW-Madison Advanced Opportunity Fellowship. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

**Institutional Review Board Statement:** Not Applicable.

**Informed Consent Statement:** Not Applicable.

**Data Availability Statement:** This paper does not report any data that has not been published elsewhere.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

