1

2 Constitutive tumor necrosis factor (TNF)-deficiency causes a reduction in spine

6 Dinko Smilovic1,2, Michael Rietsche1, Alexander Drakew1, Mario Vuksic1,2, Thomas Deller1

7
1
8 Institute of Clinical Neuroanatomy
9 Dr. Senckenberg Anatomy
10 Neuroscience Center
11 Goethe University Frankfurt
12 Theodor-Stern-Kai 7
13 D-60590 Frankfurt/Main
14
2
15 Croatian Institute for Brain Research
16 School of Medicine
17 University of Zagreb
18 Salata 12
19 10000 Zagreb, Croatia

20

21

22 Author for correspondence: Thomas Deller, M.D.

27 Keywords

28 Dentate gyrus, synaptopodin, homeostatic plasticity, synaptic plasticity, spine head, tumor necrosis factor

29

30 Acknowledgment(s)

31 We thank Charlotte Nolte-Uhl and Anke Biczysko for technical support and Dr. Meike Fellenz for help with

32 the intracellular injection method.

34 Funding/Grant sponsor

35 This work was supported by the billateral Croatian-German project (Ministry of Science and Education of

36 the Republic of Croatia and Deutsche Akademische Austauschdiesnst (MZOŠ-DAAD) to M.V. and T.D.),

37 Deutsche Forschungsgemeinschaft (DFG CRC 1080 to TD), the Scientific Centre of Excellence for Basic,

38 Clinical and Translation Neuroscience (project “Experimental and clinical research of hypoxic-ischemic

39 damage in perinatal and adult brain“; GA KK01.1.1.01.007 funded by the European Union through the

40 European Regional Development Fund) and Europass Mobility grant (to D.S).

41

42 Competing interests / Conflict of Interest Statement

43 The authors declare that the research was conducted in the absence of any commercial or financial

44 relationships that could be constructed as a potential conflict of interest.

45

46 Authors contributions

47 DS, MR, MV acquired data, DS and AD analyzed data, TD and MV conceived and supervised the study, TD,

48 DS, AD, and MV wrote the manuscript with contributions from all authors. All authors were involved in

49 data interpretation and critically revising the manuscript. All authors read and approved the final

50 manuscript.

51
52 Abstract

53 The majority of excitatory synapses terminating on cortical neurons are found on dendritic spines.

54 The geometry of spines, in particular the size of the spine head, tightly correlates with the strength of

55 the excitatory synapse formed with the spine. Under conditions of synaptic plasticity, spine geometry

56 may change, reflecting functional adaptations. Since the cytokine tumor necrosis factor (TNF) has

57 been shown to influence synaptic transmission as well as Hebbian and homeostatic forms of synaptic

58 plasticity, we speculated that TNF-deficiency may cause concomitant structural changes at the level

59 of dendritic spines. To address this question, we analyzed spine density and spine head area of Alexa-

60 568 filled granule cells in the dentate gyrus (DG) of adult C57BL/6J and TNF-deficient (TNF-KO) mice.

61 Tissue sections were double-stained for the actin-modulating and plasticity-related protein

62 Synaptopodin (SP), a molecular marker for strong and stable spines. Dendritic segments of TNF-

63 deficient granule cells exhibited ~20% fewer spines in the outer molecular layer of the dentate gyrus

64 compared to controls, indicating a reduced afferent innervation. Of note, these segments also had

65 larger spines containing larger SP-clusters. This pattern of changes is strikingly similar to the one seen

66 after denervation-associated spine loss following experimental entorhinal denervation of granule

67 cells: Denervated granule cells increase the SP-content and strength of their remaining spines to

68 homeostatically compensate for those that were lost. Our data suggest a similar compensatory

69 mechanism in TNF-deficient granule cells in response to a reduction in their afferent innervation.

70
71 Introduction

72 Dendritic spines are highly motile protrusions on dendrites of many neurons in the central nervous system

73 (Dunaevsky, Blazeski, Yuste, & Mason, 2001) forming axo-spinous synapses with glutamatergic afferents

74 (DeFelipe, Conti, van Eyck, & Manzoni, 1988). They are basic functional units of signal integration

75 detecting the coincidence of pre- and postsynaptic activity (Yuste & Denk, 1995). Of note, spines are

76 modified by synaptic activity and the geometry of spines has been tightly linked to the functional

77 properties of their synapses (Bonhoeffer & Yuste, 2002; Kasai, Matsuzaki, Noguchi, Yasumatsu, &

78 Nakahara, 2003; Matsuzaki, Honkura, Ellis-Davies, & Kasai, 2004). In particular, spine head size is

79 associated with synaptic strength (Fifková & van Harreveld, 1977; Kasai, Fukuda, Watanabe, Hayashi-

80 Takagi, & Noguchi, 2010; McKinney, 2010), post-synaptic density area (Harris, Jensen, & Tsao, 1992), and

81 AMPA-receptor density (Matsuzaki et al., 2001; Matsuzaki et al., 2004; Noguchi et al., 2011; Zito, Scheuss,

82 Knott, Hill, & Svoboda, 2009), thus making spine head size a structural surrogate marker – or at least an

83 indicator – for local synaptic activity.

84 Tumor necrosis factor (TNF) is a pleiotropic cytokine that has been implicated in a wide range of

85 physiological and pathological processes, in particular inflammation (Locksley, Killeen, & Lenardo, 2001).

86 In the central nervous system, TNF has also been shown to affect synaptic function with its net effect

87 depending strongly on its concentration (Heir & Stellwagen, 2020; Maggio & Vlachos, 2018; Santello &

88 Volterra, 2012). At low concentrations, i.e. in a physiologic range, TNF increases synaptic strength

89 (Stellwagen, Beattie, Seo, & Malenka, 2005; Stellwagen & Malenka, 2006). It is also permissive for

90 different forms of homeostatic synaptic plasticity (Becker, Zahn, Deller, & Vlachos, 2013; Steinmetz &

91 Turrigiano, 2010; Stellwagen & Malenka, 2006) and it has recently been shown to promote the ability of

92 neurons to express LTP (Maggio & Vlachos, 2018). Since TNF also reduces GABAergic transmission by

93 promoting GABAa-receptor endocytosis (Stellwagen et al., 2005), TNF has been suggested to play a role
 94 in the fine tuning of excitation/inhibition of neuronal networks (Santello & Volterra, 2012; Stellwagen et

95 al., 2005).

96 Synaptopodin (SP) is an actin-modulating protein expressed in kidney podocytes and telencephalic

97 neurons (Mundel et al., 1997). In spine-bearing neurons it is found in ~10-15% of spines, in particular in

98 the functionally important subgroup of large and stable spines (Yap et al., 2020). SP is required for the

99 formation of a spine apparatus organelle, i.e. an internal calcium store consisting of stacks of endoplasmic

100 reticulum and dense plates (Deller et al., 2003; Deller, Merten, Roth, Mundel, & Frotscher, 2000; Gray,

101 1959; Spacek, 1985). The spine apparatus promotes AMPA-R accumulation at excitatory synapses and

102 SP/spine apparatus are part of the downstream machinery changing synaptic strength (Jedlicka & Deller,

103 2017; Korkotian, Frotscher, & Segal, 2014; Vlachos et al., 2009). Accordingly, mice lacking SP show deficits

104 in Hebbian (Thomas Deller et al., 2003; Grigoryan & Segal, 2016; Jedlicka et al., 2009; Jedlicka & Deller,

105 2017; Vlachos et al., 2009; Zhang et al., 2013), homeostatic (Vlachos, Ikenberg, et al., 2013) and

106 metaplastic (Maggio & Vlachos, 2018) forms of synaptic plasticity. Recently, it has been shown that TNF

107 requires SP for its effects on Hebbian plasticity (Maggio & Vlachos, 2018), suggesting a link between TNF

108 and SP.

109 Since the number and geometry of spines are tightly linked to synaptic function, and since TNF has been

110 linked to synaptic plasticity and SP, we wondered whether lack of TNF could result in changes of spines

111 and/or SP within spines indicative of an altered network function. Our data show that constitutive TNF-

112 deficiency does indeed cause structural changes of dendrites of dentate granule cells and suggest that

113 these structural changes might reflect a compensatory homeostatic response: TNF-deficient granule cells

114 may compensate a reduced number of dendritic spines, i.e. fewer afferent synapses, by homeostatically

115 increasing the strength of some of their remaining spines, i.e. remaining synapses.

116
117 Materials and methods

118 Animals

119 Adult male mice (10-26 weeks) lacking TNF (TNF-KO, C57BL/6J Jackson Laboratory, Bar Harbor, Maine,

120 RRID: IMSR_JAX:003008; n=7) and their wildtype controls (TNF-WT, C57BL/6J background; n=7) were

121 obtained from heterozygous breeders (TNF+/-). Mice were bred and housed at the animal facility of the

122 Goethe-University Hospital Frankfurt or at MfD Diagnostics GmbH, Wendelsheim, and were maintained

123 on a 12-h light/dark cycle with food and water available ad libitum. Genotyping was performed at ~3-4

124 weeks of age. The age distribution of mice in both groups was comparable (Suppl. Fig. 1, Brunner-Munzel

125 U-test; n.s. p = 0.145). Animals were killed in accordance with the German animal welfare law and had

126 been declared to the Animal Welfare Officer of the Faculty (Wa-2014-35). Every effort was made to

127 minimize distress and pain of animals.

128

129 Intracellular injections of granule cells in fixed tissue

130 After delivery, animals were kept in an in-house scantainer for a minimum of 24 h. Animals were killed

131 with an overdose of intraperitoneal Pentobarbital and subsequently intracardially perfused (0.1 M

132 Phosphate Buffer Saline (PBS) containing 4% paraformaldehyde (PFA)). Tail biopsies were obtained after

133 death to re-confirm the genotype. Brains were taken out immediately after perfusion, post-fixed (18 h,

134 4% PFA in 0.1 M PBS, 4° C), washed trice in ice-cold 0.1 M PBS, sectioned (250µm) on a vibratome (Leica

135 VT 1000 S) and stored at 4 °C until use. Intracellular injections of granule cells in fixed slices were

136 performed as previously described (Arends & Jacquin, 1993; Hick et al., 2015; Yap et al., 2020) with

137 modifications. Hippocampal slices were placed in a custom-built, transparent, and grounded recording

138 chamber filled with ice-cold 0.1 M PBS. The chamber was attached to an epifluorescence microscope

139 (Olympus BX51WI; 10x objective LMPlanFLN10x, NA 0.25, WD 21 mm) mounted on an x-y translation table
140 (Science Products, VT-1 xy Microscope Translator). Sharp quartz-glass microelectrodes (Sutter

141 Instruments, QF100-70-10, with filament) were pulled using a P-2000 laser puller (Sutter Instruments).

142 Microelectrodes were tip-loaded with 0.75mM Alexa568-Hydrazide (Invitrogen) in HPLC-grade water

143 (VWR Chemicals, HiPerSolv CHROMANORM) and subsequently back-filled with 0.1 M LiCl in HPLC-grade

144 water. Microelectrodes were attached to an electrophoretic setup via a silver wire and 500 MΩ resistance.

145 The tip of the microelectrode was navigated into the granule cell layer using a micromanipulator

146 (Märzhäuser Wetzlar, Manipulator DC-3K). A square-wave voltage (1mV, 1 Hz) was applied using a voltage

147 generator (Gwinstek SFG-2102). Granule cells were filled under visual control for at least 10 min or until

148 no further labeling was observed. (Suppl. Fig. 2a, b). Injected sections were fixed overnight (4% PFA in

149 PBS, 4 °C, in darkness) and washed in 0.1 M PBS.

150

151 Immunohistochemistry

152 Filled and fixed injected sections (250µm) were embedded in 5% agar and resliced into 40µm thick

153 sections on a vibratome (Leica VT 1000 S). Free-floating sections were washed several times in 50 mM

154 Tris-buffered saline (TBS) containing 0.1% Triton X-100, incubated in a blocking buffer (0.5% Triton X-100,

155 5% bovine serum albumin (BSA) in 50 mM TBS) for 30 min at room temperature (RT) and subsequently

156 incubated with guinea pig anti-synaptopodin (SP, Synaptic Systems, RRID: AB_10549419; 1mg/mL, 1:2000,

157 diluted in 0.1% Triton X-100, 1%BSA in 50 mM TBS) for 3 days at RT (Paul, Choi, Schlaudraff, Deller, & Del

158 Turco, 2020). The polyclonal SP antibody recognizes AA 331 to 452 of the mouse SP protein

159 (https://sysy.com/product-factsheet/SySy_163004). After several washing steps, sections were incubated

160 with donkey anti guinea pig Alexa Fluor 488 (Jackson ImmunoResearch Labs, RRID: AB_2340472; 1mg/mL,

161 1:2000) for 4h at RT and mounted on slides (Dako fluorescence mounting medium, Dako North America

162 Inc.).
163

164 Confocal microscopy of fixed hippocampal slices

165 Confocal imaging of fixed dendritic segments from identified, Alexa568-labeled dentate granule cells in

166 the outer molecular layer (OML) of the suprapyramidal blade was done with an Olympus FV1000

167 microscope and a 60x oil-immersion objective (UPlanSApo, NA 1.35, Olympus) using FV10-ASW software

168 with 5x scan zoom at a resolution of 1024 x 1024 pixels. To ensure localization of dendritic segments in

169 the OML the hippocampal fissure was used for orientation. Dendritic segments located at a distance of

170 10-20 µm from the hippocampal fissure were identified and traced for a maximal length of 50µm. Only

171 segments within the OML were used for analysis and only 1 dendritic segment was used per labeled cell.

172 Three-dimensional image stacks of such dendritic segments (0.15 µm z-axis step size) were acquired.

173 Crossing dendritic segments or branch points were avoided to facilitate spine attribution to a given

174 segment. For imaging of dendritic segments, imaging parameters were set to capture the dendritic

175 segment as bright as possible without oversaturating the spines. For imaging of SP, all imaging parameters

176 were kept the same for all images.

177

178 Image processing and data analysis

179 Images obtained were deconvolved with Huygens Professional Version 17.10 (Scientific Volume Imaging,

180 The Netherlands, http://svi.nl). Image processing and data analysis were then performed using Fiji version

181 1.52h (Schindelin et al., 2012). Spines were identified and analyzed using established criteria (Holtmaat et

182 al. 2009). Prior to quantification, all images were randomized. Images were renamed by one experimenter

183 (Rietsche, M.) and subsequently analyzed by a second experimenter (Smilovic, D.) blind to genotypes.

184 Dendritic spines of all shapes were assessed manually on z-stacks of dendritic segments in the OML. Only

185 protrusions emanating laterally in the x-y directions, not above or below the dendrite, and exceeding the
186 dendrite for at least 5 pixels (0.2 µm) were included for analysis (Holtmaat et al., 2009; Vlachos, Müller-

187 Dahlhaus, et al., 2012; Yap et al., 2020). The length of each segment was determined (Suppl. Fig. 3a). For

188 spine head area and SP cluster area measurements, the largest maximum cross-sectional area of the spine

189 head or SP cluster in one of the x-y planes within the z-stack was manually measured using a predefined

190 gray-value as a cut-off for the border of the spine head or SP cluster (Suppl. Fig. 3b). A spine was

191 considered SP+ if the SP cluster overlapped with the spine head, neck, and/or base in both the x-z and y-

192 z directions when scrolling through the z-stacks (Suppl. Fig. 3c). The subcellular location of SP clusters in

193 the spine head, spine neck, spine base, or dendritic segment was noted. SP clusters were considered

194 within the spine head, if most (> 80%) of the SP-cluster was located within the identified area of the spine

195 head (see Suppl. Fig. 3c). SP-clusters were considered within the spine neck, if most (> 80%) of the cluster

196 was located outside the dendritic shaft border, between the identified area of the spine head and the

197 shaft, where a fluorescently filled, visible, spine neck was marked. SP clusters were considered associated

198 with the spine base, if they were found within 0.2µm of the intersection between the dendritic spine and

199 the dendritic shaft border. SP clusters were considered inside dendritic shafts if they did not meet any of

200 the aforementioned criteria but were still localized within the investigated dendritic segment.

201 Statistical analysis

202 Statistical tests and n-values used for testing are indicated in the figure captions. Statistical analysis was

203 performed using R 4.0.4 (R Core Team, 2013) called via R-script from LabVIEW scripts (National

204 Instruments, Austin, Texas). Robust methods were applied throughout as specified for the different

205 conditions: 1. Comparison of magnitude between two groups: Brunner-Munzel U-test (function

206 “brunnermunzel.permutation.test” from R-library “brunnermunzel”) (Brunner & Munzel, 2000), 2.

207 Comparison of two distributions: Cramer-von Mises test (“cvm_test” from R-library ”twosamples”), 3.

208 Comparison of proportions of occurrences between two groups: Log-likelihood ratio test (G-test) of

209 independence (“GTest” from R-library “DescTools”). In case of 2x3 tables, post hoc G-tests were applied
210 on item versus sum of the other items, 4. Comparison of proportions of occurrences against a given set of

211 proportions: G-test as goodness-of-fit test (“GTest” with given proportions) ,5. Comparisons of means

212 between groups owning levels of two crossed factors (TNF genotype and SP content of spines): robust

213 multivariate analysis of variance (MANOVA (Friedrich, Konietschke, & and Pauly, 2019)), and univariate

214 (post-hoc) tests (function “MANOVA.wide” from R-library “MANOVA.RM, considering the test statistic

215 MATS employing parametric bootstrap resampling), 6. Measure of association between two attributes:

216 Spearman´s rho (R-function “cor” employing method “spearman”), 7. Robust linear regression between

217 two attributes: R-function “rlm” employing method “MM” from library “MASS”. Confidence intervals for

218 Spearman´s rho and for regression slope and intercept were obtained using the resampling function

219 “bootstrap” from R-library “bootstrap”. In case of multiple post-hoc tests p values were adjusted using

220 the function “p.adjust” employing method “hochberg”. If P values were less than 0.05, the null-hypothesis

221 was rejected (*p < 0.05, **p < 0.01, ***p < 0.001). Quantitative data are displayed either in box plots (box

222 encloses 25% to 75% quartiles, dividing line in box represents median, x labels mean, whiskers represent

223 maximum / minimum or median + / - 1.5 times the distance between the 25% and 75% quartiles in case

224 of presence of extreme values, which are shown as additional dots) or as bar plots (mean + S.E.M.)

225 including the individual data points as dots. Diagrams were created using Microsoft Office Excel.

226
227 Results

228 Granule cell dendrites of TNF-KO mice exhibit a reduced spine density

229 Previous work showed that TNF is an important factor in the control of synaptic strength (Beattie et al.,

230 2002; Santello, Bezzi, & Volterra, 2011; Stellwagen et al., 2005; Stück et al., 2012). Since synaptic strength

231 and spine geometry are tightly linked (Bonhoeffer & Yuste, 2002; Fifková & van Harreveld, 1977; Kasai et

232 al., 2003; Matsuzaki et al., 2004), we speculated that genetic knockout of TNF in vivo may have a structural

233 correlate at the level of spines. To address this question, we first studied dendritic segments of Alexa-

234 filled granule cells (Fig. 1a-c) in the outer molecular layer of the DG of TNF-deficient and age-matched

235 C57BL/6J control mice. TNF-KO mice exhibited a significant reduction in spine density (Fig. 1d): whereas

236 wildtype mice had 2.03 spines / µm, TNF-KO mice had 1.62 spines / µm, i.e., ~20% fewer spines. Next, we

237 analyzed spine head area, since spine head area correlates well with synaptic strength and the density of

238 AMPA-Rs. Average spine head area was not significantly different between genotypes (Fig. 1e), although

239 a trend towards higher values was seen in TNF-KO segments. Finally, we calculated total spine head area

240 per segment (Fig. 1f), which illustrates how changes in spine density and head area affect the available

241 spine head area for neurotransmission. This parameter takes the number of spines into account and

242 shows that the total spine head area per segment decreases in the TNF-KO mice. After analyzing dendritic

243 segments, we shifted our attention to the entire population of spines. The trend seen in Fig. 1e was

244 confirmed to be significant (Fig. 1g). The cumulative distribution revealed a highly significant difference

245 between the two genotypes, with differences most prominent at the beginning of the curve, i.e. small

246 spine heads, and at the end of the curve, i.e. large spine heads (Fig. 1h).

247

248 TNF-KO mice show an increase in the fraction and size of large spines
249 To investigate this further, we distinguished three categories of spines (Fig. 2a): small (<0.15 µm2),

250 medium (0.15 – 0.30 µm2), and large (>0.30 µm2) sized spines and compared average spine head area

251 between control and TNF-KO spines. Although average spine head areas were not different for medium

252 sized spines, large spines were ~19% bigger and small spines were ~8% smaller in TNF-deficient granule

253 cells (Fig. 2b). These changes in spine head areas were mirrored by a significant shift in the proportions of

254 fractions of spines belonging to each of the three categories (Fig. 2c): TNF-deficient granule cells mice had

255 more small (~54% compared to ~49%) and large (~17% compared to ~13%) sized spines than controls,

256 whereas TNF-WT granule cells had more medium sized spines (38% compared to 29%) compared to TNF-

257 deficient cells.

258

259 Synaptopodin-positive spines are larger in TNF-KO mice compared to wildtype

260 Because of the conspicuous increase in the area of large spines, we wondered about the distribution of

261 the actin-modulating and plasticity-related protein SP, which is primarily associated with this subgroup of

262 spines (Thomas Deller et al., 2003; Lenz et al., 2021; Vlachos, Ikenberg, et al., 2013; Yap et al., 2020). Using

263 a double-labeling approach, Alexa-568 injected granule cells were also immunolabeled for SP (Fig. 3a, b).

264 As previously described, SP clusters were abundant in the molecular layer of the DG (Bas Orth et al., 2005;

265 Deller et al., 2000). Using single identified granule cell segments, the presence or absence of SP within

266 spines was noted and the maximum spine head area as well as the maximum cross-sectional area of SP

267 clusters were measured (Suppl. Fig. 3). In both genotypes, the majority of spines were SP- (TNF-WT, 12.8%

268 SP+; 87.2% SP-; TNF-KO, 14.0% SP+; 86.0% SP-; Fig. 3c; ***p < 0.001) and SP+ spines were significantly

269 larger than SP- spines in both genotypes (Fig. 3d; ***p < 0.001). We also noticed that SP+ spines were

270 larger in TNF-KO mice (Fig. 3d; ***p < 0.001) whereas SP- spines were smaller (Fig. 3d; *p = 0.04) (TNF-

271 WT ~0.35 µm2 SP+; ~0.15 µm2 SP-; TNF-KO ~0.45 µm2 SP+, ~0.15 µm2 SP-). Since the overall density of
272 spines is lower in TNF-deficient granule cells (Fig. 1) and the fraction of SP+ spines is constant, the absolute

273 number of SP+ spines is, however, reduced by ~14% in the TNF-KO.

274

275 Large SP+ spines are selectively enlarged in TNF-KO mice

276 We now divided SP+ and SP- spines into the three size categories (c.f. Fig. 2b). Most SP+ spines were found

277 belonging to the large spine category (TNF-WT ~53.1%, TNF-KO ~69.4%; Fig. 4a), whereas only few SP-

278 spines were in this category (TNF-WT ~6.7%; TNF-KO ~7.9%, Fig. 4a). In the subgroup of large spines, SP+

279 spines were ~23% bigger in TNF-KO mice compared to controls (Fig. 4b; ***p < 0.001) whereas large SP-

280 spines were not different between genotypes (Fig. 4b; n.s. p = 0.092). In the subgroup of small spines, SP+

281 spines were rare (TNF-WT ~12.0%, TNF-KO ~8.2%; Fig. 4a), whereas SP- spines were abundant (TNF-WT

282 ~54.5%, TNF-KO ~61.4%; Fig. 4a). SP+ spines belonging to the small category did not differ significantly

283 between TNF-deficient and TNF-WT granule cells (Fig. 4b; n.s. p = 0.593). In contrast, SP- spines showed a

284 significant reduction of ~10% in spine head area in TNF-KO mice (Fig. 4b; **p = 0.003). There was no

285 significant difference between genotypes for spines belonging to the medium sized subgroup (n.s. p =

286 0.862 for both, SP+ and SP- spines; Fig. 4b). We conclude from these findings, that (i) the increase in spine

287 head area of large spines observed in TNF-KO granule cells (Fig. 2b) is the result of an enlargement of large

288 SP+ spines (Fig. 4b), and, (ii) the reduction in spine head area of small spines (Fig. 2b) is the result of a

289 diminution of small SP- spines (Fig. 4b).

290

291 SP cluster size is increased in spines of TNF-deficient granule cells

292 The fact that SP+ spines of TNF-deficient granule cells have larger heads made us wonder whether this

293 increase is matched by a corresponding increase in SP clusters, since these two parameters are highly
294 correlated (Lenz et al., 2021; Yap et al., 2020). Indeed, average SP cluster areas were ~25% bigger in TNF-

295 deficient granule cell segments (Fig. 5a). Similarly, the cumulative distribution of SP cluster areas was

296 right-shifted in the mouse mutant compared to control (Fig. 5b). SP-clusters were preferentially found in

297 the spine head of both genotypes, without a significant shift in proportions between genotypes (Fig. 5c,

298 d). Finally, we analyzed the relationship between SP-cluster area and spine head area. Both genotypes

299 showed a strong positive correlation between the two parameters (Fig. 5e). Linear regression revealed a

300 non-significant tendency for TNF-mutants to have relatively smaller SP-clusters when comparing equally

301 sized spine heads.

303 In the present study, we analyzed the effects of constitutive TNF-deficiency on the structure of dendritic

304 spines. The rationale behind this question were earlier data implicating TNF in the modulation of synaptic

305 transmission and the fine-tuning of excitation/inhibition balance of synaptic networks (Santello

306 & Volterra, 2012; Stellwagen et al., 2005), suggesting that absence of TNF may cause changes in network

307 activity accompanied by structural alterations. Since some TNF-effects may require SP/spine apparatus

308 (Maggio & Vlachos, 2018), we also investigated SP in spines of TNF-deficient neurons. The main findings

309 of our study can be summarized as follows: (1) Granule cells of TNF-deficient mice have ~20% fewer spines

310 than wildtype controls. (2) TNF-deficient mice have an altered distribution of spine head sizes: they show

311 higher fractions of large and of small spines. (3) Although the fraction of SP+ spines was comparable

312 between genotypes, TNF-deficient mice exhibited larger SP+ spines with larger SP clusters. (4) Small SP-

313 spines were smaller in TNF-deficient mice compared to controls. This pattern of changes suggests that

314 granule cells of TNF-deficient mice have fewer afferent synapses and that the reduced number of synapses

315 is homeostatically compensated for by an increase in head size and SP cluster size of the remaining spines

316 (Fig. 6).

317

318 Structural alterations of spines in TNF-deficient mice are similar to changes observed after entorhinal

319 denervation

320 This pattern of changes, i.e. fewer spines and larger SP+ spines, is very similar to homeostatic changes

321 observed after entorhinal denervation of granule cells. In this experimental denervation model, granule

322 cell spine density is significantly reduced following entorhinal deafferentation (Caceres & Steward, 1983;

323 Vlachos, Becker, et al., 2012; Vuksic et al., 2011) and the loss of spines is homeostatically compensated

324 for by an increase in synaptic strength and an increase in SP-cluster size of the remaining spines (Vlachos,
325 Becker, et al., 2012; Vlachos, Ikenberg, et al., 2013). Such a homeostatic response may help to keep

326 denervated neurons within their physiological firing range (Platschek, Cuntz, Vuksic, Deller, & Jedlicka,

327 2016) and may promote information flow through a partially denervated brain area (Thomas Deller &

328 Frotscher, 1997; Steward, 1994).

329 In the present study focusing on TNF-deficient granule cells we recognized a comparable situation, since

330 the reduced density of spines of TNF-deficient granule cells is indicative of a reduced entorhinal

331 innervation. How this reduced innervation comes about and which developmental processes underlie this

332 change is currently unclear. Although a developmental role of TNF cannot be excluded, we consider it

333 likely that alterations in the balance of network excitation/inhibition could have caused secondary

334 changes in the density of dendritic spines. This is in line with studies linking spinogenesis and spine density

335 to afferent activity (Drakew, Müller, Gähwiler, Thompson, & Frotscher, 1996; Engert & Bonhoeffer, 1999;

336 Jourdain, Fukunaga, & Muller, 2003; Knott, Holtmaat, Wilbrecht, Welker, & Svoboda, 2006; Maletic-

337 Savatic, Malinow, & Svoboda, 1999; Nägerl, Eberhorn, Cambridge, & Bonhoeffer, 2004; Segal,

338 Greenberger, & Korkotian, 2003). Regardless of the cause, as most spines in the adult are innervated

339 (Knott et al., 2006) and excitatory input on adult spiny neurons terminates almost exclusively on spines

340 (Mates & Lund, 1983), a reduced density of spines is a bona fide structural indicator of a reduced

341 glutamatergic innervation of granule cells in the dentate gyrus of TNF-deficient mice.

342 In addition to a reduced spine density, TNF-deficient neurons show an increase in the size of large spines.

343 These spines are characterized by large PSDs (Harris et al., 1992), a high density of AMPA-R (Béïque et al.,

344 2006; Matsuzaki et al., 2001; Matsuzaki et al., 2004; Noguchi et al., 2011; Zito et al., 2009), and (this study)

345 by large SP-clusters. Thus, they are strong spines, contributing much to the excitatory drive of a neuron.

346 In contrast, many small spines lack AMPA-R and may represent "silent spines" (Kerchner & Nicoll, 2008).

347 These spines are weak spines, contributing little to the excitatory drive. Since both parameters, i.e. spine

348 head size (Matsuzaki et al., 2001; Noguchi et al., 2011; Zito et al., 2009) as well as the presence of SP-
349 clusters (Vlachos et al., 2009), are positively correlated with AMPA-R density, an increase in the spine head

350 area of SP+ spines is a bona fide structural indicator of increased synaptic strength of this spine population.

351 In sum, we see a pattern of changes in the mutants strikingly similar to what we observed under

352 experimental denervation conditions, suggesting that TNF-deficient granule cells compensate a reduction

353 in spine density - indicative of a reduced entorhinal innvervation - by homeostatically increasing the size

354 and SP content - indicative of increased synaptic strength - of their remaining spines.

355

356 TNF and SP/spine apparatus are linked

357 TNF and SP have been previously linked in the context of inflammation (Strehl et al., 2014) and Hebbian

358 plasticity (Maggio & Vlachos, 2018). In the first context, i.e. inflammation and high levels of TNF, SP was

359 found to be negatively regulated, whereas in the second context, i.e. Hebbian plasticity, TNF required SP

360 for its plasticity-promoting effect. This complex relationship may be explained by the divergent effects of

361 TNF at different concentrations: high concentrations of TNF are present in the context of inflammation,

362 whereas lower concentrations of TNF modulate synaptic plasticity under physiological conditions (Heir

363 & Stellwagen, 2020; Santello & Volterra, 2012). Thus, the relationship between TNF and SP may depend

364 on the specific condition of an experiment, i.e. whether TNF is present in high concentrations, e.g.

365 inflammation, slightly elevated concentrations, e.g. plasticity conditions, or constitutively, e.g. in naive

366 animals.

367 In this study, we used a constitutive TNF-deficient mouse and found that TNF-deficiency was associated

368 with larger SP-clusters in spines compared to wildype. A trivial explanation of this result is that TNF and

369 SP are negatively correlated. This interpretation contrasts, however, with earlier observations, in which

370 parallel increases in glial TNF (Becker et al., 2013) and SP-cluster size (Vlachos, Ikenberg, et al., 2013) were

371 reported following entorhinal denervation, suggesting that TNF and SP are positively correlated in this
372 condition. We speculate that the compensatory increase in SP cluster size seen in constitutively TNF-

373 deficient mice may not be the result of TNF-deficiency. Rather, other regulators of homeostatic synaptic

374 plasticity, such as retinoic acid (Lenz et al., 2021), which enlarges both spine heads as well as SP clusters,

375 could contribute to these changes. Further in vitro and in vivo experiments are required to resolve the

376 question, whether the increase in SP cluster sizes seen in the present study is the result of TNF-deficiency

377 or the result of denervation-related adaptations.

378

379 Competitive interactions between strong and weak spines may underlie the decrease in size of small

380 spines

381 The increased average size of large spines was accompanied by a reduced average size of small spines.

382 This reduction in size could be the result of competitive interactions between the large and strong SP+

383 spines and spines in their neighbourhood: Large SP+ spines can compete with newly formed spines via a

384 NMDA-R/CaMKII-dependent mechanism (Vlachos, Helias, Becker, Diesmann, & Deller, 2013) and may

385 reduce them in size. Such a mechanism would be homeostatic in nature, since it keeps the average spine

386 head size constant. Indeed, as was shown recently (Jungenitz et al., 2018), granule cells control spine head

387 size across their dendritic arbour. In this earlier study, high-frequency stimulation of the middle molecular

388 layer of the dentate gyrus caused LTP and an increase in spine head size in the stimulated zone. In the

389 non-stimulated outer molecular layer, however, dendritic segments of the same granule cell exhibited a

390 decrease in spine head size (Jungenitz et al., 2018), indicative of hetereosynaptic LTD in this layer

391 (Abraham & Bear, 1996; Christie & Abraham, 1992; Jedlicka, Benuskova, & Abraham, 2015). Thus, the

392 decrease in the average size of small spines may be secondary to the increase in the strength of large SP+

393 spines.

394
395 Changes of spine density and size may contribute to learning and memory deficits of TNF-deficient mice

396 The behavior of constitutive TNF-deficient mice has been studied and an impaired learning and memory

397 retention performance were reported (Baune et al., 2008). It is conceivable that the structural changes

398 we report here contribute to such behavioral deficits. A reduction of spine density by 20% is significant

399 and indicates that granule cells receive a considerably diminished entorhinal input. Furthermore, the

400 compensatory strengthening of some synapses, while keeping the neuron within a physiologic firing

401 range, could also limit the degrees of freedom a granule cell has for plasticity. Since synaptic information

402 storage capacity has been linked to spine size classes (Bromer et al., 2018), it is conceivable that

403 homostatically enlarged and highly stable SP+ spines are saturated in this respect, i.e. they cannot be

404 strengthened further and do not contribute to informational changes induced by plasticity. The

405 concomitant reduction of the fraction of medium sized spines, i.e. of those spines able to undergo both

406 structural LTP as well as structural LTD, may further limit the ability of TNF-deficient granule cells to

407 express plasticity and may thus diminish their ability to integrate into neuronal ensembles encoding novel

408 contexts (Abdou et al., 2018).

409

410
411 References
412 Abdou, K., Shehata, M., Choko, K., Nishizono, H., Matsuo, M., Muramatsu, S.‑I., & Inokuchi, K. (2018).
413 Synapse-specific representation of the identity of overlapping memory engrams. Science (New York,
414 N.Y.), 360(6394), 1227–1231. https://doi.org/10.1126/science.aat3810
415 Abraham, W. C., & Bear, M. F. (1996). Metaplasticity: the plasticity of synaptic plasticity. Trends in
416 Neurosciences, 19(4), 126–130. https://doi.org/10.1016/s0166-2236(96)80018-x
417 Arends, J. J., & Jacquin, M. F. (1993). Lucifer yellow staining in fixed brain slices: Optimal methods and
418 compatibility with somatotopic markers in neonatal brain. Journal of Neuroscience Methods, 50(3),
419 321–339. https://doi.org/10.1016/0165-0270(93)90039-t
420 Bas Orth, C., Vlachos, A., Del Turco, D., Burbach, G. J., Haas, C. A., Mundel, P. [Peter], . . . Deller, T.
421 [Thomas] (2005). Lamina-specific distribution of Synaptopodin, an actin-associated molecule essential
422 for the spine apparatus, in identified principal cell dendrites of the mouse hippocampus. The Journal
423 of Comparative Neurology, 487(3), 227–239. https://doi.org/10.1002/cne.20539
424 Baune, B. T., Wiede, F., Braun, A., Golledge, J., Arolt, V., & Koerner, H. (2008). Cognitive dysfunction in
425 mice deficient for TNF- and its receptors. American Journal of Medical Genetics. Part B,
426 Neuropsychiatric Genetics : The Official Publication of the International Society of Psychiatric Genetics,
427 147B(7), 1056–1064. https://doi.org/10.1002/ajmg.b.30712
428 Beattie, E. C., Stellwagen, D., Morishita, W., Bresnahan, J. C., Ha, B. K., Zastrow, M. von, . . . Malenka, R. C.
429 (2002). Control of synaptic strength by glial TNFalpha. Science (New York, N.Y.), 295(5563), 2282–2285.
430 https://doi.org/10.1126/science.1067859
431 Becker, D., Zahn, N., Deller, T. [Thomas], & Vlachos, A. (2013). Tumor necrosis factor alpha maintains
432 denervation-induced homeostatic synaptic plasticity of mouse dentate granule cells. Frontiers in
433 Cellular Neuroscience, 7, 257. https://doi.org/10.3389/fncel.2013.00257
434 Béïque, J.‑C., Lin, D.‑T., Kang, M.‑G., Aizawa, H., Takamiya, K., & Huganir, R. L. (2006). Synapse-specific
435 regulation of AMPA receptor function by PSD-95. Proceedings of the National Academy of Sciences of
436 the United States of America, 103(51), 19535–19540. https://doi.org/10.1073/pnas.0608492103
437 Bonhoeffer, T. [Tobias], & Yuste, R. [Rafael] (2002). Spine Motility. Neuron, 35(6), 1019–1027.
438 https://doi.org/10.1016/S0896-6273(02)00906-6
439 Bromer, C., Bartol, T. M., Bowden, J. B., Hubbard, D. D., Hanka, D. C., Gonzalez, P. V., . . . Harris, K. M.
440 [Kristen M.] (2018). Long-term potentiation expands information content of hippocampal dentate
441 gyrus synapses. Proceedings of the National Academy of Sciences of the United States of America,
442 115(10), E2410-E2418. https://doi.org/10.1073/pnas.1716189115
443 Brunner, E., & Munzel, U. (2000). The Nonparametric Behrens-Fisher Problem: Asymptotic Theory and a
444 Small-Sample Approximation. Biometrical Journal, 42(1), 17–25. https://doi.org/10.1002/(SICI)1521-
445 4036(200001)42:1<17::AID-BIMJ17>3.0.CO;2-U
446 Caceres, A., & Steward, O. [Oswald] (1983). Dendritic reorganization in the denervated dentate gyrus of
447 the rat following entorhinal cortical lesions: A Golgi and electron microscopic analysis. The Journal of
448 Comparative Neurology, 214(4), 387–403. https://doi.org/10.1002/cne.902140404
449 Christie, B. R., & Abraham, W. C. (1992). Priming of associative long-term depression in the dentate gyrus
450 by θ frequency synaptic activity. Neuron, 9(1), 79–84. https://doi.org/10.1016/0896-6273(92)90222-y
451 DeFelipe, j., Conti, F., van Eyck, S. L., & Manzoni, T. (1988). Demonstration of glutamate-positive axon
452 terminals forming asymmetric synapses in cat neocortex. Brain Research, 455(1), 162–165.
453 https://doi.org/10.1016/0006-8993(88)90127-8
454 Deller, T. [T.], Merten, T., Roth, S. U., Mundel, P. [P.], & Frotscher, M. [M.] (2000). Actin-associated protein
455 synaptopodin in the rat hippocampal formation: Localization in the spine neck and close association
456 with the spine apparatus of principal neurons. The Journal of Comparative Neurology, 418(2), 164–
457 181. https://doi.org/10.1002/(sici)1096-9861(20000306)418:2<164::aid-cne4>3.0.co;2-0
458 Deller, T. [Thomas], & Frotscher, M. [Michael] (1997). Lesion-induced plasticity of central neurons:
459 sprouting of single fibres in the rat hippocampus after unilateral entorhinal cortex lesion. Progress in
460 Neurobiology, 53(6), 687–727. https://doi.org/10.1016/s0301-0082(97)00044-0
461 Deller, T. [Thomas], Korte, M., Chabanis, S., Drakew, A. [Alexander], Schwegler, H., Stefani, G. G., . . .
462 Mundel, P. [Peter] (2003). Synaptopodin-deficient mice lack a spine apparatus and show deficits in
463 synaptic plasticity. Proceedings of the National Academy of Sciences of the United States of America,
464 100(18), 10494–10499. https://doi.org/10.1073/pnas.1832384100
465 Drakew, A. [A.], Müller, M., Gähwiler, B. H., Thompson, S. M., & Frotscher, M. [M.] (1996). Spine loss in
466 experimental epilepsy: Quantitative light and electron microscopic analysis of intracellularly stained
467 CA3 pyramidal cells in hippocampal slice cultures. Neuroscience, 70(1), 31–45.
468 https://doi.org/10.1016/0306-4522(95)00379-w
469 Dunaevsky, A., Blazeski, R., Yuste, R. [R.], & Mason, C. (2001). Spine motility with synaptic contact. Nature
470 Neuroscience, 4(7), 685–686. https://doi.org/10.1038/89460
471 Engert, F., & Bonhoeffer, T. [T.] (1999). Dendritic spine changes associated with hippocampal long-term
472 synaptic plasticity. Nature, 399(6731), 66–70. https://doi.org/10.1038/19978
473 Fifková, E., & van Harreveld, A. (1977). Long-lasting morphological changes in dendritic spines of dentate
474 granular cells following stimulation of the entorhinal area. Journal of Neurocytology, 6(2), 211–230.
475 https://doi.org/10.1007/BF01261506
476 Friedrich, S., Konietschke, F., & and Pauly, M. (2019). Resampling-Based Analysis of Multivariate Data and
477 Repeated Measures Designs with the R Package MANOVA.RM. The R Journal, 11(2), 380-400.
478 Gray, E. G. (1959). Axo-somatic and axo-dendritic synapses of the cerebral cortex: An electron microscope
479 study. Journal of Anatomy, 93(Pt 4), 420–433.
480 Grigoryan, G., & Segal, M. (2016). Ryanodine-mediated conversion of STP to LTP is lacking in
481 synaptopodin-deficient mice. Brain Structure & Function, 221(4), 2393–2397.
482 https://doi.org/10.1007/s00429-015-1026-7
483 Harris, K. M. [K. M.], Jensen, F. E., & Tsao, B. (1992). Three-dimensional structure of dendritic spines and
484 synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation
485 of synaptic physiology and long-term potentiation [published erratum appears in J Neurosci 1992
486 Aug;12(8):following table of contents]. The Journal of Neuroscience, 12(7), 2685–2705.
487 https://doi.org/10.1523/JNEUROSCI.12-07-02685.1992
488 Heir, R., & Stellwagen, D. (2020). Tnf-Mediated Homeostatic Synaptic Plasticity: From in vitro to in vivo
489 Models. Frontiers in Cellular Neuroscience, 14, 565841. https://doi.org/10.3389/fncel.2020.565841
490 Hick, M., Herrmann, U., Weyer, S. W., Mallm, J.‑P., Tschäpe, J.‑A., Borgers, M., . . . Müller, U. C. (2015).
491 Acute function of secreted amyloid precursor protein fragment APPsα in synaptic plasticity. Acta
492 Neuropathologica, 129(1), 21–37. https://doi.org/10.1007/s00401-014-1368-x
493 Holtmaat, A., Bonhoeffer, T. [Tobias], Chow, D. K., Chuckowree, J., Paola, V. de, Hofer, S. B., . . .
494 Wilbrecht, L. (2009). Long-term, high-resolution imaging in the mouse neocortex through a chronic
495 cranial window. Nature Protocols, 4(8), 1128–1144. https://doi.org/10.1038/nprot.2009.89
496 Jedlicka, P., Benuskova, L., & Abraham, W. C. (2015). Computational modeling of heterosynaptic plasticity
497 in the hippocampus. BMC Neuroscience, 16(S1). https://doi.org/10.1186/1471-2202-16-S1-P1
498 Jedlicka, P., & Deller, T. [Thomas] (2017). Understanding the role of synaptopodin and the spine apparatus
499 in Hebbian synaptic plasticity - New perspectives and the need for computational modeling.
500 Neurobiology of Learning and Memory, 138, 21–30. https://doi.org/10.1016/j.nlm.2016.07.023
501 Jedlicka, P., Schwarzacher, S. W., Winkels, R., Kienzler, F., Frotscher, M. [Michael], Bramham, C. R., . . .
502 Deller, T. [Thomas] (2009). Impairment of in vivo theta-burst long-term potentiation and network
503 excitability in the dentate gyrus of synaptopodin-deficient mice lacking the spine apparatus and the
504 cisternal organelle. Hippocampus, 19(2), 130–140. https://doi.org/10.1002/hipo.20489
505 Jourdain, P., Fukunaga, K., & Muller, D. (2003). Calcium/Calmodulin-Dependent Protein Kinase II
506 Contributes to Activity-Dependent Filopodia Growth and Spine Formation. The Journal of
507 Neuroscience, 23(33), 10645–10649. https://doi.org/10.1523/JNEUROSCI.23-33-10645.2003
508 Jungenitz, T., Beining, M., Radic, T., Deller, T. [Thomas], Cuntz, H., Jedlicka, P., & Schwarzacher, S. W.
509 (2018). Structural homo- and heterosynaptic plasticity in mature and adult newborn rat hippocampal
510 granule cells. Proceedings of the National Academy of Sciences of the United States of America,
511 115(20), E4670-E4679. https://doi.org/10.1073/pnas.1801889115
512 Kasai, H. [Haruo], Fukuda, M., Watanabe, S., Hayashi-Takagi, A., & Noguchi, J. (2010). Structural dynamics
513 of dendritic spines in memory and cognition. Trends in Neurosciences, 33(3), 121–129.
514 https://doi.org/10.1016/j.tins.2010.01.001
515 Kasai, H. [Haruo], Matsuzaki, M. [Masanori], Noguchi, J., Yasumatsu, N., & Nakahara, H. (2003).
516 Structure–stability–function relationships of dendritic spines. Trends in Neurosciences, 26(7), 360–368.
517 https://doi.org/10.1016/S0166-2236(03)00162-0
518 Kerchner, G. A., & Nicoll, R. A. (2008). Silent synapses and the emergence of a postsynaptic mechanism
519 for LTP. Nature Reviews. Neuroscience, 9(11), 813–825. https://doi.org/10.1038/nrn2501
520 Knott, G. W., Holtmaat, A., Wilbrecht, L., Welker, E., & Svoboda, K. [Karel] (2006). Spine growth precedes
521 synapse formation in the adult neocortex in vivo. Nature Neuroscience, 9(9), 1117–1124.
522 https://doi.org/10.1038/nn1747
523 Korkotian, E., Frotscher, M. [Michael], & Segal, M. (2014). Synaptopodin regulates spine plasticity:
524 Mediation by calcium stores. The Journal of Neuroscience : The Official Journal of the Society for
525 Neuroscience, 34(35), 11641–11651. https://doi.org/10.1523/JNEUROSCI.0381-14.2014
526 Lenz, M., Kruse, P., Eichler, A., Muellerleile, J., Straehle, J., Jedlicka, P., . . . Vlachos, A. (2021). All-Trans
527 Retinoic Acid induces synaptic plasticity in human cortical neurons. ELife, in press.
528 Locksley, R. M., Killeen, N., & Lenardo, M. J. (2001). The TNF and TNF Receptor Superfamilies. Cell, 104(4),
529 487–501. https://doi.org/10.1016/s0092-8674(01)00237-9
530 Maggio, N., & Vlachos, A. (2018). Tumor necrosis factor (TNF) modulates synaptic plasticity in a
531 concentration-dependent manner through intracellular calcium stores. Journal of Molecular Medicine
532 (Berlin, Germany), 96(10), 1039–1047. https://doi.org/10.1007/s00109-018-1674-1
533 Maletic-Savatic, M., Malinow, R., & Svoboda, K. [K.] (1999). Rapid dendritic morphogenesis in CA1
534 hippocampal dendrites induced by synaptic activity. Science (New York, N.Y.), 283(5409), 1923–1927.
535 https://doi.org/10.1126/science.283.5409.1923
536 Mates, S. L., & Lund, J. S. (1983). Neuronal composition and development in lamina 4C of monkey striate
537 cortex. The Journal of Comparative Neurology, 221(1), 60–90. https://doi.org/10.1002/cne.902210106
538 Matsuzaki, M. [M.], Ellis-Davies, G. C., Nemoto, T., Miyashita, Y., Iino, M., & Kasai, H. [H.] (2001). Dendritic
539 spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons.
540 Nature Neuroscience, 4(11), 1086–1092. https://doi.org/10.1038/nn736
541 Matsuzaki, M. [Masanori], Honkura, N., Ellis-Davies, G. C. R., & Kasai, H. [Haruo] (2004). Structural basis
542 of long-term potentiation in single dendritic spines. Nature, 429(6993), 761–766.
543 https://doi.org/10.1038/nature02617
544 McKinney, R. A. (2010). Excitatory amino acid involvement in dendritic spine formation, maintenance and
545 remodelling. The Journal of Physiology, 588(Pt 1), 107–116.
546 https://doi.org/10.1113/jphysiol.2009.178905
547 Mundel, P. [P.], Heid, H. W., Mundel, T. M., Krüger, M., Reiser, J., & Kriz, W. (1997). Synaptopodin: An
548 actin-associated protein in telencephalic dendrites and renal podocytes. The Journal of Cell Biology,
549 139(1), 193–204. https://doi.org/10.1083/jcb.139.1.193
550 Nägerl, U. V., Eberhorn, N., Cambridge, S. B., & Bonhoeffer, T. [Tobias] (2004). Bidirectional activity-
551 dependent morphological plasticity in hippocampal neurons. Neuron, 44(5), 759–767.
552 https://doi.org/10.1016/j.neuron.2004.11.016
553 Noguchi, J., Nagaoka, A., Watanabe, S., Ellis-Davies, G. C. R., Kitamura, K., Kano, M., . . . Kasai, H. [Haruo]
554 (2011). In vivo two-photon uncaging of glutamate revealing the structure-function relationships of
555 dendritic spines in the neocortex of adult mice. The Journal of Physiology, 589(Pt 10), 2447–2457.
556 https://doi.org/10.1113/jphysiol.2011.207100
557 Paul, M. H., Choi, M., Schlaudraff, J., Deller, T. [Thomas], & Del Turco, D. (2020). Granule Cell Ensembles
558 in Mouse Dentate Gyrus Rapidly Upregulate the Plasticity-Related Protein Synaptopodin after
559 Exploration Behavior. Cerebral Cortex (New York, N.Y. : 1991), 30(4), 2185–2198.
560 https://doi.org/10.1093/cercor/bhz231
561 Platschek, S., Cuntz, H., Vuksic, M., Deller, T. [Thomas], & Jedlicka, P. (2016). A general homeostatic
562 principle following lesion induced dendritic remodeling. Acta Neuropathologica Communications, 4,
563 19. https://doi.org/10.1186/s40478-016-0285-8
564 R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical
565 Computing, Vienna, Austria. URL http://www.R-project.org/.
566 Santello, M., Bezzi, P., & Volterra, A. (2011). Tnfα controls glutamatergic gliotransmission in the
567 hippocampal dentate gyrus. Neuron, 69(5), 988–1001. https://doi.org/10.1016/j.neuron.2011.02.003
568 Santello, M., & Volterra, A. (2012). Tnfα in synaptic function: Switching gears. Trends in Neurosciences,
569 35(10), 638–647. https://doi.org/10.1016/j.tins.2012.06.001
570 Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., . . . Cardona, A. (2012).
571 Fiji: An open-source platform for biological-image analysis. Nature Methods, 9(7), 676–682.
572 https://doi.org/10.1038/nmeth.2019
573 Segal, M., Greenberger, V., & Korkotian, E. (2003). Formation of dendritic spines in cultured striatal
574 neurons depends on excitatory afferent activity. The European Journal of Neuroscience, 17(12), 2573–
575 2585. https://doi.org/10.1046/j.1460-9568.2003.02696.x
576 Spacek, J. (1985). Three-dimensional analysis of dendritic spines. Ii. Spine apparatus and other cytoplasmic
577 components. Anatomy and Embryology, 171(2), 235–243. https://doi.org/10.1007/bf00341418
578 Steinmetz, C. C., & Turrigiano, G. G. (2010). Tumor necrosis factor-α signaling maintains the ability of
579 cortical synapses to express synaptic scaling. The Journal of Neuroscience : The Official Journal of the
580 Society for Neuroscience, 30(44), 14685–14690. https://doi.org/10.1523/JNEUROSCI.2210-10.2010
581 Stellwagen, D., Beattie, E. C., Seo, J. Y., & Malenka, R. C. (2005). Differential regulation of AMPA receptor
582 and GABA receptor trafficking by tumor necrosis factor-alpha. The Journal of Neuroscience : The Official
583 Journal of the Society for Neuroscience, 25(12), 3219–3228. https://doi.org/10.1523/JNEUROSCI.4486-
584 04.2005
585 Stellwagen, D., & Malenka, R. C. (2006). Synaptic scaling mediated by glial TNF-alpha. Nature, 440(7087),
586 1054–1059. https://doi.org/10.1038/nature04671
587 Steward, O. [O.] (1994). “Reorganization of neuronal circuitry following central nervous system trauma:
588 naturally occurring processes and opportunities for therapeutic intervention”: The Neurobiology of
589 Central Nervous System Trauma. New York, Oxford: Oxford University Press.
590 Strehl, A., Lenz, M., Itsekson-Hayosh, Z., Becker, D., Chapman, J., Deller, T. [Thomas], . . . Vlachos, A.
591 (2014). Systemic inflammation is associated with a reduction in Synaptopodin expression in the mouse
592 hippocampus. Experimental Neurology, 261, 230–235.
593 https://doi.org/10.1016/j.expneurol.2014.04.033
594 Stück, E. D., Christensen, R. N., Huie, J. R., Tovar, C. A., Miller, B. A., Nout, Y. S., . . . Ferguson, A. R. (2012).
595 Tumor necrosis factor alpha mediates GABA(A) receptor trafficking to the plasma membrane of spinal
596 cord neurons in vivo. Neural Plasticity, 2012, 261345. https://doi.org/10.1155/2012/261345
597 Vlachos, A., Becker, D., Jedlicka, P., Winkels, R., Roeper, J., & Deller, T. [Thomas] (2012). Entorhinal
598 denervation induces homeostatic synaptic scaling of excitatory postsynapses of dentate granule cells
599 in mouse organotypic slice cultures. PloS One, 7(3), e32883.
600 https://doi.org/10.1371/journal.pone.0032883
601 Vlachos, A., Helias, M., Becker, D., Diesmann, M., & Deller, T. [Thomas] (2013). Nmda-receptor inhibition
602 increases spine stability of denervated mouse dentate granule cells and accelerates spine density
603 recovery following entorhinal denervation in vitro. Neurobiology of Disease, 59, 267–276.
604 https://doi.org/10.1016/j.nbd.2013.07.018
605 Vlachos, A., Ikenberg, B., Lenz, M., Becker, D., Reifenberg, K., Bas-Orth, C., & Deller, T. [Thomas] (2013).
606 Synaptopodin regulates denervation-induced homeostatic synaptic plasticity. Proceedings of the
607 National Academy of Sciences of the United States of America, 110(20), 8242–8247.
608 https://doi.org/10.1073/pnas.1213677110
609 Vlachos, A., Korkotian, E., Schonfeld, E., Copanaki, E., Deller, T. [Thomas], & Segal, M. (2009).
610 Synaptopodin regulates plasticity of dendritic spines in hippocampal neurons. The Journal of
611 Neuroscience, 29(4), 1017–1033. https://doi.org/10.1523/JNEUROSCI.5528-08.2009
612 Vlachos, A., Müller-Dahlhaus, F., Rosskopp, J., Lenz, M., Ziemann, U., & Deller, T. [Thomas] (2012).
613 Repetitive magnetic stimulation induces functional and structural plasticity of excitatory postsynapses
614 in mouse organotypic hippocampal slice cultures. The Journal of Neuroscience, 32(48), 17514–17523.
615 https://doi.org/10.1523/JNEUROSCI.0409-12.2012
616 Vuksic, M., Del Turco, D., Vlachos, A., Schuldt, G., Müller, C. M., Schneider, G., & Deller, T. [Thomas]
617 (2011). Unilateral entorhinal denervation leads to long-lasting dendritic alterations of mouse
618 hippocampal granule cells. Experimental Neurology, 230(2), 176–185.
619 https://doi.org/10.1016/j.expneurol.2011.04.011
620 Yap, K., Drakew, A. [Alexander], Smilovic, D., Rietsche, M., Paul, M. H., Vuksic, M., . . . Deller, T. [Thomas]
621 (2020). The actin-modulating protein synaptopodin mediates long-term survival of dendritic spines.
622 ELife, 9. https://doi.org/10.7554/eLife.62944
623 Yuste, R. [R.], & Denk, W. (1995). Dendritic spines as basic functional units of neuronal integration. Nature,
624 375(6533), 682–684. https://doi.org/10.1038/375682a0
625 Zhang, X., Pöschel, B., Faul, C., Upreti, C., Stanton, P. K., & Mundel, P. [Peter] (2013). Essential role for
626 synaptopodin in dendritic spine plasticity of the developing hippocampus. The Journal of Neuroscience
627 : The Official Journal of the Society for Neuroscience, 33(30), 12510–12518.
628 https://doi.org/10.1523/JNEUROSCI.2983-12.2013
629 Zito, K., Scheuss, V., Knott, G., Hill, T., & Svoboda, K. [Karel] (2009). Rapid functional maturation of nascent
630 dendritic spines. Neuron, 61(2), 247–258. https://doi.org/10.1016/j.neuron.2008.10.054
631
632 Figures and figure captions:

633
634 Figure 1 Granule cell dendrites of TNF-KO mice exhibit a reduced spine density. (a) Granule cell in the
635 suprapyramidal blade of the dentate gyrus of a TNF-alpha-knock-out (TNF-KO) mouse intracellularly filled
636 with the fluorescent dye Alexa-568 hydrazide (magenta; fixed tissue). Dendritic segments in the outer
637 molecular layer (OML) were used for analysis. MML: middle molecular layer. IML: inner molecular layer.
638 GCL: granule cell layer. HF: hippocampal fissure. Scale bar = 20 µm. (b, c) Single confocal sections of
639 granule cell dendrites in the OML of TNF-WT (b) and TNF-KO (c) mice. Scale bar = 2 um. Large (thick arrow),
640 medium (thin arrow), and small (arrowhead) spines are indicated. (d) Density of dendritic spines on TNF-
641 KO segments is significantly reduced (~1.62 1/µm) compared to controls (~2.03 1/µm). Analysis based on
642 7 WT and 7 KO mice; 1 segment per cell; 3 dendritic segments per animal (n=21 segments); 1632 TNF-WT;
643 1303 TNF-KO spines. *p = 0.0123, Brunner-Munzel U-test. (e) Average spine head area per segment of
644 TNF-WT (~0.18 µm2) and TNF-KO (~0.20 µm2) mice are not significantly (n.s.) different; p = 0.991; Brunner-
645 Munzel U-test; n = 21 segments per group. (f) Total spine head area of all spines of a segment divided by
646 the length of the analyzed segment shows a ~14% reduction in spine area per µm of the segment for TNF-
647 KO (~0.31 µm2/µm) compared to TNF-WT (~0.36 µm2/µm), *p = 0.0385; Brunner-Munzel U-test. n = 21
648 segments per group. (g) Box plots of spine head area of all spines from all segments. Spine heads of TNF-
649 KO mice (0.190 µm²) are larger than spine heads of TNF-WT mice (0.178 µm²); 1632 TNF-WT; 1303 TNF-
650 KO spines. *p = 0.0311, Brunner-Munzel U-test. (h) Cumulative frequency plot of these spine head areas.
651 Distributions are different between genotypes; ***p < 0.001; Cramer-von Mises test.

652
653

654 Figure 2 TNF-KO mice show an increase in the fraction and area of large spines. (a) Spines were divided
655 into three classes: small (<0.15 µm2), medium (0.15 – 0.30 µm2), and large (>0.30 µm2) spines. Scale bar =
656 0.5 µm. (b) Box plots of spine head area of all spines from all segments allocated to the three size classes.
657 Large spines were ~19% bigger in TNF-KO mice (~0.52 µm2) compared to TNF-WT mice (~0.43 µm2); ***p
658 < 0.001; while small sized TNF-KO spines (~0.0815 µm2) were ~8% smaller compared to TNF-WT spines
659 (~0.0885 µm2); **p = 0.001. Medium sized spines were not significantly (n.s.) different. p = 0.637;
660 univariate robust MANOVA separately for the three size classes; TNF-WT n = 800, 625, 207; TNF-KO n =
661 703; 385; 215 (small/medium/large). (c) Proportions of spine numbers allocated to the three size classes
662 were different between genotypes; ***p < 0.001; G-test of independence. TNF-KO mice had more large
663 (~17% compared to ~13%; **p = 0.008; post hoc G-test) and more small (~54% compared to ~49%; **p =
664 0.007; post hoc G-test) spines and fewer medium spines (~30% compared to ~38%; ***p < 0.001; post
665 hoc G-test) compared to WT.

666
667

668 Figure 3 Synaptopodin+ spines are larger in TNF-KO mice compared to wildtype. Alexa-568 injected
669 granule cells of wildtype (a) and TNF-KO (b) mice were double-labeled for Synaptopodin (SP; green).
670 Arrows point to SP-positive (SP+) spines; Arrowheads mark SP- negative (SP-) spines. Scale bars = 1 µm.
671 (c) Proportions of fractions of SP+ and SP- spines are not different between genotypes; n.s. p = 0.328; G-
672 test of independence. Proportion of fractions of SP- in TNF-WT (n = 1423, ~87.2%) and in TNF-KO (n =
673 1120, ~86.0%) and fractions of SP+ (n = 209, ~12.8%) spines in TNF-WT and SP+ (n = 183, ~14.0%) spines
674 in TNF-KO animals differed from equality (50% for SP+ and SP-, respectively); ***p < 0.001, G-test of
675 goodness of fit. (d) Box plots of head area of SP+ and SP- spines of TNF-WT and TNF-KO animals. (TNF-WT
676 ~0.35 µm2 SP+ (n = 209); ~0.15 µm2 SP- (n = 1423); TNF-KO ~0.45 µm2 SP+ (n = 183), ~0.15 µm2 SP- (n =
677 1120)). SP+ spines have larger spine head areas compared to SP- spines in both genotypes. ***p < 0.001;
678 univariate post hoc tests following robust MANOVA. SP+ spines were larger in TNF-KO mice (***p < 0.001),
679 whereas SP- spines were smaller in TNF-KO mice, *p = 0.04; univariate post-hoc tests following robust
680 MANOVA with p-values adjusted for multiple comparisons.

681
682

683 Figure 4 Large SP+ spines are selectively enlarged in TNF-KO mice. (a) Proportions of fractions of spines
684 allocated to the three size classes seperately for SP- (left panel, blue in the middle panel) and SP+ (right
685 hand panel, red in the middle panel) spines. For both, SP- and SP+ spines, the proportions were different
686 between TNF-WT and TNF-KO mice; ***p < 0.001 for SP- and **p = 0.004 for SP+; G-test of independence.
687 In both genotypes, the majority of SP- spines were small spines and the majority of SP+ spines were large
688 spines. Of note, the fraction of large SP+ spines (**p = 0.003; post hoc G-test) as well as the fraction of
689 small SP- spines (***p < 0.001; post hoc G-test) were larger in TNF-KO mice, whereas the fraction of
690 medium SP- spines (***p < 0.001; post hoc G-test) as well as medium SP+ spines (*p = 0.0123; post hoc
691 G-test) was smaller in TNF-KO mice. Fractions of large SP- (p = 0.284; post hoc G-test) as well as small SP+
692 spines (p = 0.216; post hoc G-test) did not differ significantly between genotypes. Numbers of occurences:
693 TNF-WT: SP- spines n = 1423 total; n = 775/552/96 small/medium/large SP- spines; SP+ spines n = 209
694 total; n = 25/73/111 small/medium/large SP+ spines; TNF-KO: SP- spines n = 1120 total; 688/344/88
695 small/medium/large SP- spines; SP+ spines n = 183 total; n = 15/41/127 small/medium/large SP+ spines.
696 (b) Box plots of head area of SP+ and SP- spines of TNF-WT and TNF-KO animals separately for the three
697 size classes. TNF-KO mice have bigger large sized SP+ spines (~0.58 µm2) compared to WT (~0.46 µm2)
698 (***p < 0.001) and smaller small sized SP- spines (~0.08 µm2) compared to WT (~0.09 µm2) (**p = 0.003).
699 Differences between other groups were not significant (n.s.): Small SP+ spines (WT: ~0.111 µm², KO:
700 ~0.116 µm². p = 0.593); medium SP- (WT: ~0.205 µm², KO: ~0.205 µm². p = 0.862); medium SP+ (WT:
701 ~0.227 µm², KO: ~0.221 µm². p = 0.862); large SP- (WT: ~0.386 µm², KO: 0.413 µm². p = 0.092); post-hoc
702 univariate pairwise comparisons following robust MANOVA with p-values adjusted for multiple
703 comparisons.

704
705

706 Figure 5 SP cluster size is increased in spines of TNF-deficient granule cells. (a) Box plots of Area of SP
707 clusters in SP+ spines. SP clusters were ~33% larger in TNF-KO mice (~0.13 µm², n = 183) compared to TNF-
708 WT (~0.10 µm², n = 209) spines. **p = 0.0018; Brunner-Munzel U-test. (b) Cumulative frequency plot of
709 SP cluster areas for both genotypes. ***p < 0.001, Cramer-von Mises test. (c) Higher magnifications of
710 dendritic segments of a TNF-WT mouse immunolabeled for SP. SP clusters were found in the head, neck,
711 or base of spines (arrows). Scale bars = 1 µm. (d) Localization of SP clusters in TNF-WT and TNF-KO mice.
712 TNF-WT: 53.1/27.8/19.1%, head/neck/base; TNF-KO: 57.4/24.0/18.6%, head/neck/base. TNF-WT n = 209;
713 TNF-KO = 183 clusters; n.s. p = 0.652. G-test of independence. (e) SP cluster area is positively correlated
714 to spine head area in both genotypes. TNF-WT: 99% confidence interval of Spearman’s ρ = 0.306 to 0.60;
715 TNF-KO: 99% confidence interval of Spearman’s ρ = 0.153 to 0.511. Lines represent robust linear
716 regressions of SP cluster area on spine head size. Regression slopes were not significantly different
717 between genotypes, i.e. confidence intervals do overlap: TNF-WT: 0.208, 95% confidence interval = 0.144
718 to 0.263 TNF-KO: 0.107, 95% confidence interval = 0.064 to 0.148.

719
720

721 Figure 6 Summary diagram illustrating the structural differences between dendritic segments of TNF-
722 WT and TNF-KO granule cells (a, b) TNF-KO dendrites exhibit ~20% fewer spines compared to controls.
723 Although average spine head area is not significantly different, the size distribution of spines has changed:
724 Compared to controls, the fraction of large spines as well as the average size of large spines is increased
725 in the KO. The fraction of small spines is also increased, but this group of spines showed a decreased
726 average head size. Thus, the KO exhibits relatively more large spines, which are also larger and at the same
727 time more small spines, which are also smaller. These bidirectional changes explain why average spine
728 head size between genotypes is not significantly different. Furthermore, large spines were found to be
729 associated with the plasticity-related protein Synaptopodin (green clusters). The larger spines of TNF-KO
730 mice also exhibited larger SP clusters.

731
732

733 Supplementary figure 1. Age distribution of TNF-WT and TNF-KO mice. Mice used in the present study
734 were age-matched littermates. There were no significant differences between the age distribution of the
735 two groups. n.s. p = 0.144. Brunner-Munzel U-test. TNF-WT, n=7; TNF-KO, n=7.

736
737

738 Supplementary figure 2. Intracellular injections of granule cells in fixed slices. (a) Schematic showing
739 intracellular filling of dentate granule cells in a frontal 250 µm-section of a mouse hippocampus. A quartz
740 electrode loaded with 1 mM Alexa568 dye and backfilled with 0.1 M LiCl2 is used to impale and
741 subsequently fill (10 min) neurons. DGsupra: suprapyramidal blade of Dentate Gyrus. DGinfra: infrapyramidal
742 blade of dentate gyrus. CA1, CA2, CA3: Areas 1, 2 and 3 of Cornu Ammonis. Scale bar = 500 µm. (b)
743 Overview image showing several dye-injected granule cells in the DGsupra. Scale bar = 100 µm.

744
745
746 Supplementary figure 3. Quantification of dendritic spine head area and synaptopodin (SP) cluster area.
747 (a) Maximum projection of a confocal image stack showing a dendritic segment of a granule cell. For
748 analysis, dendritic spines of all shapes were assessed manually on z-stacks of dendritic segments. Only
749 protrusions extending laterally in the x-y plane, not above or below the dendrite, and exceeding the
750 dendrite for at least 5 pixels (0.2 µm) were included for analysis. For spine head area the maximum cross-
751 sectional area of the spine head was identified in one of the x-y planes and measured. Scale bar = 5 µm.
752 (b) Only spines with heads protruding at least 0.2 µm from the parent dendrite (parallel white lines) were
753 analyzed. Spines co-localizing with an SP cluster in the x-y, x-z, and y-z directions when scrolling through
754 the z-stack were considered to be SP+ (arrow). Scale bar = 0.5 µm. (c) Spine head area and SP cluster area
755 were defined as the largest x-y cross-sectional area obtained in a z-stack. Images containing the largest
756 area of spine head (middle column, orange outline of spine head, asterisk) and SP cluster (right column,
757 orange outline of SP-cluster, asterisk) are highlighted. Scale bar = 0.5 µm.