TNFAIP8 family gene expressions in the mouse tail intervertebral disc injury model

Abstract Introduction The TNF‐α‐induced protein‐8 (TNFAIP8, also known as TIPE) family of molecules comprises four members: TNFAIP8 and TIPEs1‐3. Since the first description of these proteins, their roles in fine‐tuning inflammation and in directing leukocyte migration have been described in several organ systems. However, their relationship with intervertebral disc (IVD) is unknown. Materials and methods Here, we describe the expression of TNFAIP8 family genes in the nucleus pulposus (NP) and annulus fibrosus (AF) of the normal adult murine IVD. We further describe the expression of these genes in the injured male and female murine IVD. Results Tnfaip8 gene expression was decreased, and Tipe1 gene expression was essentially unchanged, in response to injury. Tipe2 and Tipe3 gene expression was markedly elevated in response to IVD injury, along with those encoding known inflammatory markers (ie, Tnfa, Il6, Cxcl1, and Adam8). Additionally, sex‐related differences were also observed for some of these genes in intact and injured mouse IVDs. Future studies include examining tissue distribution of TNFAIP8 family proteins and identifying cells that produce them. In addition, examining mice that are deficient in TNFAIP8 molecules, in relation to gene expression, tissue morphology and mouse behavior, may further delineate the roles of these molecules in IVD inflammation and degeneration.


| INTRODUCTION
Back pain related to intervertebral disc (IVD) degeneration costs billions of dollars in the U.S. 1,2 Current treatments include surgical and nonsurgical approaches, and often result in incomplete symptom relief because the molecular relationships between back pain and IVD degeneration are still unclear. Clinically, patients with chronic back pain often have histories of injuries. Uncovering mechanisms of disc inflammation following injury may shed light on the mechanism of spinal pain.
The TNFAIP8 (tumor necrosis factor-α-induced protein 8; TNFAIP8) family are newly described regulators of immunity and tumorigenesis consisting of four highly homologous mammalian proteins: TNFAIP8, TIPE1 (TNFAIP8-like 1, or TNFAIP8L1), TIPE2 (TNFAIP8L2), and TIPE3 (TNFAIP8L3). They are novel signaling proteins recently described to be key factors regulating inflammation and oncogenesis. [3][4][5] Interestingly, despite significant sequence homology among the four members of this family, 5 they are involved in different biological activities and exhibit remarkable variability of expression. 6 Furthermore, this family of proteins is highly dysregulated in arthritis, cancers and various other chronic diseases. 6 TNFAIP8 is a founding member of the TNFAIP8 family, originally identified in head and neck cancer cells, 7 and in cells treated with TNF-α. 8 TNFAIP8 is a negative regulator of apoptosis, and is oncogenic. 4,5 Moreover, TNFAIP8 protein inhibits apoptosis by inhibiting Caspase-8. 9 Knocking out the Tnfaip8 gene exacerbated disease in a dextran sodium sulfate (DSS) model of murine colitis. 10 Indeed, phospho-AKT (also known as protein kinase B [PKB]) activity decreased in TNFAIP8 knockout colonic cells, indicating that TNFAIP8 may regulate epithelial cell death by targeting AKT. 11 The increased mortality in TNFAIP8 knockout mice with colitis could be explained by increased cell death, and decreased proliferation of colonic epithelial cells. 11 In mice, TIPE1 was found to be expressed in a wide variety of tissues and cells, including neurons, hepatocytes, muscle tissue, intestinal epithelial cells, and germ cells. TIPE1 interacts with Rac1 and inhibits activation of downstream p65 and c-Jun N-terminal kinase signaling, which increased caspase-mediated apoptosis. 12 TIPE2 was discovered by genomic profiling of inflamed neural tissues. 10,13 TIPE2 was found to be widely expressed in lymphoid tissues; its knockout in mice resulted in the development of spontaneous multiorgan inflammation, splenomegaly and premature death, and these mice were hypersensitive to toll-like receptor (TLR) stimulation. 10 Recently, it was found that leukocyte polarity was generated by TIPE2, a transfer protein for phosphoinositide second messengers.
TIPE2 functioned as a local enhancer of phosphoinositide-dependent signaling and cytoskeleton remodeling, which promoted leading-edge formation. 14 TIPE3 is the most recently investigated TNFAIP8 family member.
TIPE3 protein is preferentially expressed in reproductive and neural tissues, with nearly identical murine and human expression profiles. 15 TIPE3 is the transfer protein of lipid second messengers that promote cancer. 16 Several TIPE proteins have been crystallized. 17,18 Knockout mice for TNFAIP8 and TIPE2 have also been generated, 10,11 and should be valuable for studying inflammation in the musculoskeletal system, including back pain related to IVD injuries. Inflammatory cytokines and chemokines have been found in painful/degenerative human IVDs. [19][20][21] However, TNFAIP8 family gene expression in mouse IVDs has not been described previously. This study is the first that describes gene expression profiles of the TIPE molecules in response to injury, an important step toward understanding the roles of the TNFAIP8 family of proteins in IVD inflammation.
The mouse model permits examination of IVD degeneration in mice with genetic modifications, allowing the role of a specific molecule in the disease process to be studied. Overall, mice and humans share virtually the same set of genes. Almost every gene found in one species so far has been found in a closely related form in the other. It is, however, important to be aware of the limitations of modeling human spinal disease using animals, 22 and to recognize that the biomechanical properties of the mouse tail differ from those of the human spine. The mouse tail IVD injury model was established by Yang et al 23 and Martin et al. 24 We have recently refined the procedure using a percutaneous needle puncture approach, and have shown that needle injury results in a reproducible course of morphological and molecular changes consistent with IVD degeneration. 25,26 Among the four time points we studied, 26 reproducible changes in gene expression and histology occurred at 1 week post injury. Accordingly, this time point was selected to examine TNFAIP8 family member gene expression.
DBA mice are used in the musculoskeletal field, because they are more susceptible to some types of arthritis than the C57BL/6 (also known as B6) strain. 27 We have found that ADAM8 plays important roles in arthritis [28][29][30] and in IVD degeneration. 31 Our ongoing work on the role of Adam8 in intervertebral disc degeneration uses a mouse with ADAM8 inactivated by introducing the E330Q mutation into its proteolytic domain. 30 This animal is on the DBA background, and the Tnfaip8 gene expression studies are performed with these mice.
Since TNFα is a known inducer of the TNFAIP8 family, the Tnfα gene was also included in the panel of genes examined. We have also examined genes encoding inflammatory cytokines/chemokines related disc degeneration such as Il6 20 and Cxcl1 21 in this study. The Adam8 gene was included because of its relationship with inflammation and fibronectin cleavage in the IVD. 31 Finally, in order to comment on potential differences in gene expression pattern between male and female animals per National Institutes of Health (NIH) guidelines, we have examined previously mice of both sexes and found subtle but at times significant differences in expression patterns of selected genes. 32 Therefore, we have analyzed data from male and female mice separately in the present study. Harbor, Maine) were used in this study. Twelve female and 12 male mice were used for tail IVD injury experiment, and 6 female mice were used to isolate nucleus pulposus (NP) and annulus fibrosus (AF) tissues for RNA extraction. Mice were housed under pathogenfree conditions with environmental enrichment (nestlets by Ancare, Bellmore, New York), and up to five animals per cage. Mice were fed PicoLab diet no. 5053 (LabDiet, Fort Worth, Texas) without restriction, provided with acidified bottled water, and maintained on a 12:12-hours light:dark cycle. Room temperature is at 21.1 C to 24.4 C (equivalent to 70 F-76 F) and humidity at 30% to 70%.

| NP and AF dissection
The lumbar and coccygeal vertebrae were isolated with a scalpel under a dissecting microscope (VistaVision, VWR International, Radnor, Pennsylvania), as described previously. 33 The gelatinous NP was scraped off with a scalpel. AF tissues, identified by their concentric rings, were shaved off the cartilaginous endplate with a scalpel.
Lumbar and coccygeal IVDs were pooled for each animal.

| Tail injury surgery
Surgery was performed as described previously. 26 Specifically, each mouse was anesthetized with Ketamine (90 mg/kg) and Xylazine (10 mg/kg) subcutaneously. Under anesthesia, the skin was cleaned with betadine. Under fluoroscopic guidance with a mini C-arm (OrthoScan FD Pulse Mini C-Arm, Orthoscan Inc., Scottsdale, AZ), the mouse coccygeal (Co) IVDs were identified, and a 26G needle was inserted into the IVD space until the needle tip reached 2/3 of the disc thickness. This information has been included in our recent manuscript as a supplemental figure. 26 Care was taken not to puncture the opposing wall of the AF. Indeed, when the opposing AF wall was damaged, a more severe injury was seen on MRI. 25 Gelatinous tissues were often found on the needle tip after this had been removed, suggesting that the needle puncture induces an acute herniation of the gelatinous NP of the IVD. In the current study, the Co3/4 and Co5/6 IVDs in each mouse were injured, while Co4/5 and Co6/7 served as intact controls, as described previously. 26 Mice were checked 4 hours after surgery and the next day. The mice were then monitored daily until the endpoint. Even though mice did not show any signs of pain or distress, as a pre-emptive treatment, mice were given buprenorphine at the time of sedation and 4 hours later.
Animals were sacrificed by exposure to CO 2 at 1 week after tail disc injury. From each mouse tail, Co3/4 (injured) and Co4/5 (intact control) discs were isolated individually for RNA extraction. Specifically, IVD tissues were separated from their adjacent cartilaginous endplates and bone with a scalpel, under a dissecting microscope (VistaVision, VWR International, Radnor, Pennsylvania). Co5/6 (injured) and Co6/7 (intact control) were isolated en bloc for histological examination.

| RNA isolation and quantitative real-time PCR
Total cellular RNA was isolated by the Trizol method as described previously. 26,33 Specifically, the isolated IVD tissues were soaked in RNALater

| Histological evaluation
The IVDs and portions of the adjacent bony vertebral bodies were isolated immediately after euthanasia. The disc with its surrounding vertebral bodies was fixed with 4% paraformaldehyde for 24 hours. The bone-discbone segments were decalcified with a solution consisting of 12.5% EDTA for 1 week, with shaking, until the bony portion was completely decalcified. 26 The tissues were then dehydrated and embedded in paraffin and sectioned to 5 μM thickness. For Picrosirius red staining, sections were stained with 0.1% Picrosirius red (Sigma) for 45 minutes. All samples were examined under a light microscope (Nikon) and photographed.

| Statistics
The expression ratio for genes of interest (Tnfaip8, Tipe1-3, Il6, Cxcl1, Adam8, and Tnfα) and house-keeping gene (Gapdh) were calculated for each sample. If expression ratio was not normally distributed, data were log transformed. To assess differences in gene expression ratio, two-way analysis of variance (ANOVA) in repeated measures where the grouping factor was sex, and the repeated measure injury/intact were performed. To minimize type I error, post hoc pairwise comparisons using Tukey-Kramer tests were performed to examine differences between injured/intact tissues and tissues from mice of different sexes. A P-value of <.05 was considered statistically significant. All analyses were performed using SAS statistical software (Version 9.4, SAS Institute, Cary, North Carolina). 37 3 | RESULTS

| Tnfaip8 family gene expression in the intact NP and AF tissues
To examine whether Tnfaip8 family of genes is expressed in the intact mouse IVDs, we used mice that have not had tail IVD injuries. We isolated NP and AF tissues as described previously. 33 The expression ratio for genes of interest (Tnfaip8, and Tipe1-3) and housekeeping gene (Gapdh) were calculated for NP and AF. There was detectable expression of all four members of the TNFAIP8 family, although the overall levels were low compared with those of the extracellular matrix molecules. 33 There was no statistically significant difference in tnfaip8 gene expression between NP and AF (Tnfaip8 to Gapdh gene expression ratio: 0.0346 and 0.0540 in the NP and AF, respectively; n = 6 mice, P = .2049; Figure 1A). Tipe1 was expressed at a higher level in the AF than in the NP (Tipe1 to Gapdh ratio 0.0068 and 0.0493 in the NP and AF, respectively; n = 6; P = .0312; Figure 1B). The gene expression level of Tipe2 was very low in both NP and AF tissues (Tipe2 to Gapdh ratio: 1.27 × 10 −7 and 1.44 × 10 −5 in the NP and AF, respectively; n = 6, P = .1181; Figure 1C). It is worth mentioning that these levels are very low and quantification may not be reliable with the real-time PCR methods used here. Tipe3 gene showed a tendency to higher expression in the AF than in the NP (Tipe3 to Gapdh ratio 0.0036 and 0.0297 in the NP and AF, respectively; n = 6; P = .0582; Figure 1D). Among the four genes, Tnfaip8 gene expression is the highest, and Tipe2 gene expression lowest both in NP and AF, with Tnfaip8 gene expression 5 and 3 orders of magnitude of Tipe2 in NP and AF, respectively (n = 6).

| Tnfaip8 family gene expression in the injured tail IVDs
In male mice, Tnfaip8 gene expression decreased in response to injury, compared with intact controls (Tnfaip8 to Gapdh gene expression ratio: 0.0068 in injured IVDs and 0.0128 in controls respectively; n = 12 mice; P = .0017; Figure 2A). In female mice, however, there was no statistically significant difference in Tnfaip8 gene expression in injured and intact IVDs (Tnfaip8 to Gapdh gene expression ratio: Similarly, Tnfaip8 gene expression was higher in female injured than male injured discs (P = .0003; Figure 2A).
There was no statistically significant change in Tipe1 gene expression in injured IVDs compared with that in intact controls either in male mice (Tipe1 to Gapdh gene expression ratio: 0.0088 in injured, 0.0109 in controls respectively; n = 12 mice; P = .4525) or in female mice (Tipe1 to Gapdh gene expression ratio: 0.0255 in injured, 0.0180 in controls respectively; n = 11 mice; P = .7232). When analyzing all male and female mice together, Tipe1 gene expression in injured vs intact controls still did not show any statistical significance (n = 23, P = .2032). Similar to the Tnfaip8 gene, Tipe1 showed a trend for higher gene expression in female intact control discs than in male controls (P = .0599). The gene expression ratios in the injured IVDs are higher in injured females than males (P = .0274; Figure 2B).
Tipe2 gene expression, on average, is three orders of magnitude lower in the intact IVDs compared with that of Tnfaip8 (n = 23 mice).
Tipe2 gene expression increased significantly in mice of both sexes in response to injury, compared with intact controls. In male mice, Tipe2 to Gapdh gene expression ratios were 60.67 × 10 −6 in injured discs and 5.55 × 10 −6 in controls, respectively (n = 12 mice; P = .0048). In female mice, Tipe2 gene expression also increased in injured com- in injured IVDs and 339 × 10 −5 in controls, respectively; n = 11, P < .0001). This interesting pattern of change in response to injury is due, at least in part, to higher Tipe3 gene expression in female control IVDs than in male controls (P < .0001; Figure 2D).  Figure 3A). Surprisingly, the males expressed higher levels of Tnfα than did female mice, both in injured discs and controls (P = .0004 and P = .0005 for injured and control discs, respectively; Figure 3A).

| Proinflammatory cytokine (Tnfα, Il6, Cxcl1) and Adam8 gene expression
The increase in Il6 gene expression did not reach statistical significance in either male or female mice when comparing injured discs with intact controls (n = 15 mice/group; P =.9818 and .3085, respectively). When data from both male and female mice are analyzed together, Il6 gene expression elevation in injured IVDs compared with controls reached statistical significance (Il6 to Gapdh gene expression ratio: 6.79 × 10 −4 in injured discs, 2.71 × 10 −4 in controls; n = 30 mice; P = .0437). There was no statistically significant difference between the male and female mice, either in injured discs or controls (n = 15 mice/group; P > .0500; Figure 3B).  Figure 3D). There was no statistically significant difference between the male and female mice, either in injured discs or in controls (n = 15 mice/group; P > .0500; Figure 3D).

| Tail IVDs showed consistent histological changes following injury in mice of both sexes
We have examined histological features of the mouse tails following injury in mice on the DBA background, with Picrosirius red, to reveal changes in the content and orientation of the collagen bundles. There F I G U R E 3 Pro-inflammatory gene expression in the injured mouse tail intervertebral disc. A, Tnfα gene expression; B, Il6 gene expression; C, Cxcl1 gene expression; D, Adam8 gene expression. Each point shows data from one intervertebral disc. **P ≤ .01; *P ≤ .05 were consistent histological changes in both male and female IVDs post injury (n = 3-9; Figure 4). Specifically, normal NP and AF architecture is revealed by Picrosirius red staining, in both the male and female mice ( Figure 4A-A 00 , C-C 00 ). Following injury, the IVD tissue showed loss of normal NP architecture, with collagen-rich scar tissue replacing the normal NP ( Figure 4B 0 ,D 0 ). Annular rings were distorted with some interruptions following injury ( Figures 1D 00 and 4B). Consistent with prior findings, 32 we did not detect any significant differences in histological features between the male and female mice.

| DISCUSSION
The TNFAIP8 family consists of four members, and recently has been found to regulate inflammatory processes, by interacting with inflammatory pathways and directing lymphocyte migration. 4,14 TNFAI8 and TIPE2-deficient mice have been generated, 10,11 and may prove useful in studying inflammatory processes such as back pain and osteoarthritis in the musculoskeletal field. In the present manuscript, we described gene expression in intact and injured IVDs, which serves as baseline information for future mechanistic studies aimed at explaining potential roles of TNFAIP8 family members in homeostasis of the normal discs and their response to injury.
In the intact IVDs, we were able to separate the NP and AF tissues based on gelatinous texture of the NP and presence of the annular rings in the AF. Gene expression of the four Tnfaip8 family members was analyzed separately (Figure 1). Similar to the profile of extracellular matrix and adhesion molecule genes, 33 the TNFAIP8 family of genes tend to be more highly expressed in the AF than in NP. Among the four family members, the Tnfaip8 gene expression level is the highest, and Tipe2 gene expression the lowest, with its level barely measurable. This finding suggests that Tnfaip8 may be expressed by native disc tissues, and possibly important for normal tissue maintenance. Tipe2, known to be expressed mainly by leukocytes, 4 may perform immune functions. This division of labor by molecules of the same family is reminiscent of the two subtypes of cyclooxygenases. 38 The NP and AF tissues could no longer be reliably separated 1 week following needle puncture ( Figure 4). Therefore, the tissues were analyzed as whole IVDs. In response to injury, Tnfaip8 gene expression decreased while Tipe2 and Tipe3 gene expression increased ( Figure 2). These findings suggest that TNFAIP8 family members may perform different functions in the injury/repair process. TIPE2, preferentially found in leukocytes, may be expressed by infiltrating macrophages in response to injury, which has been shown in human degenerating IVDs. 20,39 Confirming or refuting this hypothesis is an important future direction. Future work could include in situ hybridization to show the distribution of mRNAs, and immunostaining to reveal the distribution of TNFAIP8 family protein and leukocyte markers, to identify the cell types that produce these molecules. One limitation of the study is that we have found subtle differences in biomechanical properties between bone-disc-bone motion segments of the mouse tail (Zhang-Mauck, unpublished   However, including both sexes in experimental design would increase the animal number needed to design experiments with adequate power for data analysis, in order to yield unbiased and reproducible results. We have recently found and reported subtle differences in extracellular matrix and inflammatory gene expression between male and female mice. 32 In the current study, we have found that Tnfaip8 and Tipe3 gene expression is significantly higher in female control than in their male counterparts (Figure 2A and D), while Tnfα gene expression is higher in both intact and injured male discs than in female tissues ( Figure 3A). These findings, together with our previous findings, 32 confirm the need for studies in both male and female animals in the disc injury model. TNFα treatment has been shown to alter F-actin and alpha-tubulin in AF cells, suggestive of altered cytoskeletal stiffness. 40 Anti-TNFα (infliximab) significantly inhibited pro-inflammatory cytokine production, while anti-IL6 did not. TNFα altered AF cell mechanobiology with cytoskeletal remodeling that potentially sensitized AF cells to mechanical strain and increased TNFα-induced pro-inflammatory cytokine production. 40 Exposure to TNFα induced expression of additional pro-inflammatory cytokines and altered IVD mechanical behavior. 41 One potential mechanism of TNFα-induced cytoskeleton change may be by altering the function of the TNFAIP8 family members, since they have been found recently to direct leukocyte migration by reorganizing the cytoskeleton. 14 Examining leukocyte infiltration in the injured IVDs would be an important future direction, to determine the relative contribution of inflammatory mediators by resident disc cells and infiltrating leukocytes.
TNFα has been shown in herniated and degenerative human IVD tissues. [42][43][44] The TNFα receptor (TNFR) has two subtypes, TNFR1 and TNFR2, with sometimes opposing functions to maintain the delicate balance of inflammation, cell death, and proliferation. 45,46 Interestingly, the level of TNFR1 has been positively correlated with back and radicular leg pain, while that of TNFR2 is negatively correlated with pain. 44 Injecting TNFα into the lumbar IVD induces pain behavior and disc degeneration in rats. 47 However, some clinical studies have shown efficacy of reducing back and radicular pain by protein factors that regulate TNFα signaling, 48-50 but other such work did not show any benefit. 51 The inconsistencies in clinical outcomes may be due, in part, to the fact that protein factors require injections, which was delivered either into the epidural space or subcutaneous tissue in the above-mentioned trials. Permeability into the disc is limited by either method. Furthermore, these reagents block both TNFR1 and TNFR2, and may result in differing effects depending on disease stage, inciting factors, and other individual patient-related factors such as genetics.
The present work includes a study of IL6, a known marker of IVD degeneration. 20,52 CXCL1 and ADAM8, subjects of our previous studies, 21,26,31,53 have also been included. All three genes were found to be elevated in both male and female mice post injury, confirming that they are reliable molecular markers for disc injury. A broader range of molecules warrants examination. For example, we have examined differences in expression of 84 genes between NP and AF. 33 A similar strategy could be used to examine differences in sex and genetic background in response to injury, in the wild type and TNFAIP8-deficient mice. Novel methodology, such as RNASeq, could also be utilized to examine an even larger number of genes.
In the aging mouse, the chondrocyte-like NP cells have been shown to be of notochordal lineage. 54 Although using lineage-tracing method is not feasible in humans, it would not be surprising if the human NP cells after age 10 to 12 years of age are also of notochordal origin, although the cell morphology differs from that of notochordal cells in young humans. In addition, all four members of the mammalian TNFAIP8 family are highly conserved in their amino acid sequences. 5 Further studies in other species are therefore indicated, since it is highly likely that the role of TNFAIP8 family members in local inflammation is translatable to other mammalian species including humans.
In conclusion, genes of the Tnfaip8 family are expressed in the intact mouse IVDs. Tipe2 and Tipe3 genes are elevated in response to injury, along with other known inflammatory markers (ie, Tnfα, il6, Cxcl1, and Adam8). Surprisingly, the Tnfaip8 gene was downregulated