*2.4. Hu Proteins Expression in Mouse DRG Neurons Is Decreased in Diabetic and Diabetic Resistant Mice Compared to Control Mice*

The Hu protein expression in mouse DRG neurons was evaluated by immunofluorescence in control (Figure 5A–C), diabetic resistant (Figure 5D–F), and diabetic (Figure 5G–I) mice. We evidenced the expression of HuB, HuC, and HuD proteins in DRG neurons for all three CD-1 mice groups. We localized all three Hu proteins both in the soma and the neurites of the DRG neurons. Each Hu protein has a distinct distribution pattern in the soma and particularly HuC tends to organize in clusters. We observe a pronounced localisation of HuB and HuC in the neurites in diabetic conditions compared to diabetic resistant or control conditions.

**Figure 5.** Hu proteins (HuB, HuC, and HuD ) expression in DRG neurons of control (**A**–**C**), diabetic resistant (**D**–**F**), and diabetic mice (**G**–**I**). The red labeling is obtained with rabbit polyclonal anti-ELAVL2, anti-ELAVL3, and anti-ELAVL4 antibodies, respectively, followed by the staining with donkey polyclonal anti-rabbit conjugated with Rhodamine Red X. Images are captured with an LSM 710 Zeiss laser scanning microscope using a 63× oil objective. Scale bar 10 μm.

Further on, we performed the quantitative analysis of the neuronal Hu proteins expression based on the mean fluorescence intensity (Figure 6) and we correlated these results with the *Elav*-like gene expression. In control mice, we obtained the following ranking for the protein expression HuD > HuC > HuB, and the one-way ANOVA analysis followed by post-hoc Bonferroni test indicated statistical significance between HuD and HuB expression (*p* < 0.001) and between HuD and HuC expression (*p* < 0.001). However, the distinct levels of HuB, HuC, and HuD expression in control mice are not in agreement with the *Elav*-like gene levels that are comparable. In diabetic and diabetic resistant conditions neuronal HuB, HuC, and HuD proteins were distinctly regulated compared to control conditions. The two-way ANOVA analysis indicated statistical significance of Hu proteins expression, for the diabetic condition and for their interaction (Table S11). In comparison with control mice, HuB protein was significanly downregulated in diabetic resistant mice and upregulated in diabetic mice, HuC protein was significantly downregulated in diabetic resistant mice, and HuD protein was significantly downregulated both in diabetic and diabetic resistant mice (Table S12).

**Figure 6.** Hu protein expression based on mean fluorescence intensity analysis in DRG neurons of control, diabetic resistant, and diabetic mice. Data are expressed as mean ± SD in the captured images. Statistical significance is indicated \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

### **3. Discussion**

In this study, we brought evidence that Hu proteins undergo expression changes that might be associated with the diabetic condition. First of all, it is necessary to discuss the model of diabetes that we employed in our study. Indeed, several mouse models for type 1 diabetes have been developed, the most employed being the STZ-induced diabetes, despite its variability, depending on the mice strain [34] or the development of diabetic neuropathy [35]. To detail, variable concentrations of STZ (single i.p. injection) were used in different mice strains to induce diabetes, i.e., ICR, ddY and BALB/c: 100–200 mg/kg and C57BL/6: 75–150 mg/kg [34]. Considering the different mice strain sensitivity to STZ-induction of diabetogenic state [35–37], in our study, we decided to induce diabetes in CD-1 adult mice with single i.p. STZ injection (150 mg/kg).

Despite its variability the STZ-induced model of diabetes is robust and used in multiple studies. However, researchers focus either on the mortality rate or on the resistance of the animal strain when injected with STZ, but in the STZ-"sensitive" animal strains little attention is paid to the rate of surviving animals that are resistant to the STZ-induction of diabetes. Some studies reported a subpopulation of mice [37] or rats [38] remains normoglycemic upon STZ-induction of diabetes, but do not explicitly consider these animals as "diabetic resistant". In our study, we are classifying STZ-resistant CD1-mice as "diabetic resistant" mice.

The resistance to STZ-induction of diabetes was previously described in different mice strains, including mice lacking phosphatase and a tensin homolog deleted from chromosome 10 [39] or nonobese diabetes-resistant mice [40,41]. On the other hand, the resistance to STZ-induction of diabetes in a certain percentage of animals belonging to the so-called 'sensitive' strains (e.g., CD-1 mice) is generally not discussed. For example, in CD-1 mice, the distinction between the induction of type 1 diabetes by single injection of high STZ dose (130 mg/kg or 150 mg/kg) and the induction of type 2 diabetes by multiple injections of low STZ dose (40 mg/kg) was reported [42], but the percentage of 100% reported in the text for the induction of diabetes animals by single injection of high STZ dose is not in agreement with the percentage of animals with hyperglycemia ≥600 mg/dL, one out of five animals (130 mg/kg STZ, 16% mortality) and three out of three (150 mg/kg STZ, 50% mortality) presented in the same study. In our study, we demonstrate the resistance to the induction of diabetes

with a single injection of high STZ dose (150 mg/kg) in approximately 46% (CD-1 adult male mice, six out of 13 mice) of the surviving animals (37% mortality). We also evidence that diabetic-resistant mice have lower body weight values compared to control mice, but remain normoglycemic. Our data, indicating that diabetic mice have a lower body weight and a higher glycemia compared to control mice, are in agreement with previous protocols of STZ-induction of diabetes [37]. To resume, our study demonstrates that, in addition to the expected population of STZ-induced hyperglycemic CD-1 mice, a mice subpopulation develops resistance to STZ and we consider that a distinct analysis of the STZ-resistant normoglycemic mice should be done in each study, as the findings might be relevant in understanding the mechanisms of insulin resistance development in human patients.

In diabetic patients, diabetic peripheral neuropathy is gradually characterized by hyperalgesia and allodynia, followed by the development of hypoalgesia and finally the complete loss of sensation [43]. STZ-induced diabetes in mice or rat is associated with thermal hyperalgesia in early phases [43] and with thermal hypoalgesia in late stages of diabetes [44–46] in the absence of insulin therapy. Commonly thermal hypoalgesia precedes epidermal denervation in STZ-diabetic mice [47]. We confirm an increase in the paw withdrawal latency (thermal hypoalgesia) in diabetic mice 8-weeks after STZ-induction of diabetes, while diabetic resistant mice have similar paw withdrawal latencies compared to control mice. The absence of changes in the algesic profile of diabetic resistant mice is supported by previous reports showing that ~50% male Sprague-Dawley rats remain normoglycemic after STZ-injection, without significant changes in the algesic profile (no changes in the threshold or latency to heat noxious stimuli, or in the pressure pain threshold and frequency of withdrawal to brush and 20-g von Frey filament) compared to control rats [38].

Considering previous studies that reported the role played by Hu proteins from DRG neurons in hyperalgesia [19,48–52], we analyzed the expression changes of Hu proteins in diabetic and diabetic resistant mice compared to control mice and correlated these data with algesic profile to radiant heat exposure. To detail, HuR contributes to hyperalgesia either associated with experimental autoimmune encephalomyelitis [48] or with inflammation (exposure to bradykinin and interleukin-1) where stabilized cyclooxygenase-2 mRNA [49]. HuD is upregulated and contributes to pain hypersensitivity to mechanical and cold stimulation in antiretroviral-evoked painful neuropathy by regulating spinal ryanodine receptor-2 [50] or GAP43 [19,51] or contributes to thermal hot hyperalgesia in oxaliplatin-induced neuropathy by regulating GAP43 [52]. On the other hand, to the best of our knowledge, this is the first study documenting the role of Hu proteins in hypoalgesia associated with the diabetic condition, addressing the expression of Hu proteins in DRG neurons. Considering the previous reports regarding the role of *nELAVL* Hu-proteins in neuronal excitability by binding to the mRNAs encoding proteins from the glutamate synthesis pathway [15], or encoding Kv1.1 voltage-gated potassium channels [16], we might suppose that *nELAVL* Hu-proteins might also stabilize / regulate mRNA encoding other proteins (i.e., ion channels) involved in DRG neuronal excitability and being important players in the algesic profile and diabetes.

Our study brings evidence that *Elavl* genes and Hu proteins expression is distinctly regulated in DRG sensory neurons in diabetic, diabetic resistant, and control conditions, and we have tried to correlate these expression data with the final paw withdrawal latency in the hot plate test. Interestingly, the final paw withdrawal latency in diabetic-resistant mice has not significantly changed in comparison to control mice, which indicates that diabetic-resistance mice do not undergo changes in the algesic profile in radiant heat exposure after 8 weeks. In diabetic mice, *Elavl2* and *Elavl3* are downregulated, while HuB is upregulated and HuD is downregulated, compared to control mice. In diabetic resistant mice, both *Elavl* genes and Hu proteins are strongly downregulated, compared to control mice. It is very interesting to remark that, despite the lack of changes in the algesic profile of diabetic resistant mice, we reported significant *Elavl* gene and Hu protein expression changes in diabetic resistant mice compared to diabetic or control mice. Previous studies indicated HuD upregulation in thermal hyperalgesia [19,50–52] and our study brings evidence that HuD is downregulated in thermal hypoalgesia induced by the advanced diabetes status. Considering the role played by HuD upregulation

in nerve regeneration upon lesion [21], a possible scenario in diabetes would be: (i) hyperalgesia (early phases of diabetes) is associated with HuD upregulation involved in nerve regeneration, (ii) hypoalgesia (late phases of diabetes) is associated with HuD downregulation, when its ability to regulate mRNA proteins involved in nerve recovery is overcome. However, HuD downregulation in late diabetes should be considered with caution as STZ-induced the same kind of expression changes in normoglycemic diabetic resistant mice. Extensive analysis of the algesic profile in diabetic and diabetic resistant mice in correlation with Hu proteins expression is necessary.

Our study also analyzed the immunolocalization of Hu proteins in correlation with the diabetic status. Previous immunostaining studies documented the expression of HuD [8,53], HuC/HuD (anti-16A11 antibody) [17], or all Hu proteins (anti-16A11 antibody) [18] in adult DRG neurons. The HuD immunopositivity in DRGs neurons was analyzed: (i) in the cell compartments, with distribution both in the soma and the axons [19] or (ii) in the subcellular structures (strong staining in the cytoplasm [8,18] and low staining in the nucleus, Golgi apparatus and mitochondria [18]). Our study indicates HuB, HuC, and HuD expression in the soma and neurites of the DRG neurons. However, our semi-quantitative analysis of Hu protein expression was limited to the soma of DRG neurons.

Although specific immunostaining was obtained for all three neuronal Hu proteins in different structures of the central nervous system or in the spinal cord [54], only HuD specific immunopositivity was analyzed in DRGs [8,53], but no specific targeting of HuB and HuC expression in DRGs was done. In our study, we bring evidence of the specific localisation of HuB and HuC in DRG neurons, and we also demonstrate that all three Hu proteins undergo expression changes in late diabetes.

Different neuronal types from hippocampus, cerebellum, olfactory cortex, neocortex, etc. were demonstrated to express from one to several Hu genes [54]. We might suppose that different subtypes of DRG neurons express various combinations of Hu genes, distinctly contributing to the regulation/stabilization of mRNA encoding proteins involved in the development of diabetic neuropathy and/or thermal hypoalgesia. To this purpose, subsequent colocalization studies of Hu proteins in DRG neurons might bring new insights.

To resume, our study analyzed the distinction between diabetic and diabetic resistant mice in the STZ-induction model and compares them with the control mice. We correlate the diabetic state with hyperglycemia, lower body weight, presence of late thermal hypoalgesia, *Elavl2* and *Elavl3* downregulation, HuB upregulation, and HuD downregulation in comparison to control conditions. Meanwhile, we correlate the diabetic resistant state with normoglycemia, slightly lower body weight, normal algesia, strong *Elavl2, Elavl3,* and *Elavl4* downregulation, HuB, HuC, and HuD downregulation compared to control conditions. In conclusion, we demonstrate the distinct expression regulation of *nELAVL* Hu proteins in diabetes and we consider that it is very important to understand if these Hu protein expression changes are also present in patients with peripheral diabetic neuropathy and if there is any correlation with the status of the disease.

#### **4. Materials and Methods**

## *4.1. Animals*

Adult CD-1 male mice aged 6 weeks with a mean body weight of 20 g were acquired from the "Cantacuzino" Medico-Military National Institute of Research and Development. Animals (*N* = 40) were housed 3/cage in the animal husbandry of 'Horia Hulubei' National Institute of Physics and Nuclear Engineering, with food and water *ad libitum*. All procedures were in accordance with the European Guidelines on Laboratory Animal Care, and with the approval of the institutional Ethics Committee of the 'Horia Hulubei' National Institute of Physics and Nuclear Engineering (approval number 31/11.06.2015).
