Human INCL fibroblasts display abnormal mitochondrial and lysosomal networks and heightened susceptibility to ROS-induced cell death

Infantile Neuronal Ceroid Lipofuscinosis (INCL) is a pediatric neurodegenerative disorder characterized by progressive retinal and central nervous system deterioration during infancy. This lysosomal storage disorder results from a deficiency in the Palmitoyl Protein Thioesterase 1 (PPT1) enzyme - a lysosomal hydrolase which cleaves fatty acid chains such as palmitate from lipid-modified proteins. In the absence of PPT1 activity, these proteins fail to be degraded, leading to the accumulation of autofluorescence storage material in the lysosome. The underlying molecular mechanisms leading to INCL pathology remain poorly understood. A role for oxidative stress has been postulated, yet little evidence has been reported to support this possibility. Here we present a comprehensive cellular characterization of human PPT1-deficient fibroblast cells harboring Met1Ile and Tyr247His compound heterozygous mutations. We detected autofluorescence storage material and observed distinct organellar abnormalities of the lysosomal and mitochondrial structures, which supported previous postulations about the role of ER, mitochondria and oxidative stress in INCL. An increase in the number of lysosomal structures was found in INCL patient fibroblasts, which suggested an upregulation of lysosomal biogenesis, and an association with endoplasmic reticulum stress response. The mitochondrial network also displayed abnormal spherical punctate morphology instead of normal elongated tubules with extensive branching, supporting the involvement of mitochondrial and oxidative stress in INCL cell death. Autofluorescence accumulation and lysosomal pathologies can be mitigated in the presence of conditioned wild type media suggesting that a partial restoration via passive introduction of the enzyme into the cellular environment may be possible. We also demonstrated, for the first time, that human INCL fibroblasts have a heightened susceptibility to exogenous reactive oxygen species (ROS)-induced cell death, which suggested an elevated basal level of endogenous ROS in the mutant cell. Collectively, these findings support the role of intracellular organellar networks in INCL pathology, possibly due to oxidative stress.

Human INCL fibroblasts display abnormal mitochondrial and lysosomal networks and 7 heightened susceptibility to ROS-induced cell death 8 9 10 11 Bailey Balouch 1,3 , Halle Nagorsky 1 , Truc Pham 2,4 , Thai LaGraff 2,5 , and Quynh Chu-LaGraff 1,2* 12 13 Email: chulagrq@union.edu 31 network also displayed abnormal spherical punctate morphology instead of normal elongated 48 tubules with extensive branching, supporting the involvement of mitochondrial and oxidative 49 stress in INCL cell death. Autofluorescence accumulation and lysosomal pathologies can be 50 mitigated in the presence of conditioned wild type media suggesting that a partial restoration via 51 passive introduction of the enzyme into the cellular environment may be possible. We also 52 demonstrated, for the first time, that human INCL fibroblasts have a heightened susceptibility to 53 exogenous reactive oxygen species (ROS)-induced cell death, which suggested an elevated basal 54 Introduction lipofuscin storage materials [4,9,13,16,19,20]. The accumulation of ceroid or lipofuscin in 81 lysosomes is characteristic of all subtypes of Batten Disease [21] and is heterogeneous in 82 composition, consisting of proteins, proteolipids and metals [19,20]. Specifically in INCL 83 neurons, these lipid-protein aggregates appear in the form of granular osmiophilic deposits 84 (GRODs) and are curvilinear, fingerprint, or rectilinear shaped [13,21,22] as detected by 85 electron microscopy studies [9,20,23]. GRODs have been identified in neurons as well as non-86 neuronal cell types including lymphocytes [23,24], fibroblasts [23,25,26], and brown adipose 87 tissues [12]. 88 The underlying pathology of INCL and how PPT1 enzyme deficiency leads to neuronal 89 cell death remains relatively not well understood [17]. Oxidative stress and related damage is a 90 common pathological feature of numerous neurodegenerative disorders [27,28]. Studies using 91 human INCL brains and PPT1 knock-out mice revealed that the loss of PPT1 leads to caspase 92 activated pathway of apoptosis in neurons, presumably due to ER-induced stress responses [10, 93 16]. Excess storage material from the lysosome may be trafficked back to the ER, activating the 94 unfolded protein response (UPR), causing ER stress [10,17,19]. Reactive oxygen species (ROS) 95 are released from the ER in response to stress, triggering mitochondrial-mediated apoptosis, and 96 contributing to neurodegeneration [17,19,29]. Neurons exhibit elevated energetic needs and 97 thus depend heavily on oxidative metabolism and produce higher levels of ROS than other cell 98 types, increasing their susceptibility to oxidative stress [29,30]. Additionally, while PPT1 is 99 localized to lysosomes in all cell types; in neurons, it is also present in synaptic vesicles 100 facilitating the recycling of synaptic vesicles after the release of neurotransmitters. PPT1 101 deficiency in neurons causes reduced availability of synaptic vesicles at axon terminals, possibly 102 contributing to the progressive neurodegeneration observed in INCL [4,19]. 103 Glutamine (Life Technologies, California, USA). Cultures were incubated at 37° C and 5% CO2. 141 Cultures of two human fibroblast lines, wild type GM05659 (WT) and INCL patient 142 GM20389 (PT), were used to produce the WT-and PT-conditioned media. Conditioned media 143 was obtained by saving media from either WT or PT two-three days old cultures. Cells were 144 maintained at confluency in T25 flasks for six weeks in order to prevent proliferation and allow 145 for better modeling of the post-mitotic state of neurons. Media was replaced every 2-3 days with 146 50% fresh media and 50% of either WT-or PT-conditioned media. Four conditioned groups 147 were created -group 1: WT culture receiving 50% WT conditioned media (WT+WT), group 2: 148 WT culture receiving 50% PT conditioned media (WT+PT), group 3: PT culture receiving 50% 149 WT media (PT+WT), and group 4: PT culture receiving 50% PT conditioned media (PT+PT). primary antibodies for 60 minutes, and secondary antibodies for 45 minutes. All incubations took 171 place at 37° C and 5% CO2. Staining was followed by three additional PBS washes, and a final 172 wash in dH2O. 173

Fluorescence Imaging and Analysis 174
Following fixation and staining with the appropriate antibodies or intracellular markers, 175 coverslips were mounted onto microscope slides using anti-fade medium (Molecular Probes-176 Invitrogen, Oregon, USA), and visualized with a Zeiss AX10 Observer A1 inverted microscope, 177 equipped with a SPOT imaging camera and software (Diagnostic Instruments Inc, Michigan, 178 USA). Fluorescence was observed using DAPI (ex 358nm / em 461nm), GFP (ex 488nm / em 179 530nm), and Texas Red (ex 596nm / em 620nm) filter channels. Cells were viewed with 100X-180 oil immersion and 40X objectives. 181 For quantitative fluorescence imaging, all cells were imaged at 100X magnification with 182 oil immersion objective. The imaging parameters were optimized for every staining. Slides free 183 of previous fluorescence exposure were imaged using identical parameters. To account for 184 photobleaching, all slides were imaged sequentially, and fluorescence exposure was timed and 185 limited to less than 45 minutes per slide. Only non-overlapping cells were imaged for 186 quantitative fluorescence analysis, in order to not mistake combined signal for increased signal 187 intensity. Images were analyzed using the freehand selection tool in ImageJ (NIH, Maryland, 188

USA) to trace the perimeter of cells and obtain measurements of mean and maximum intensity. 189
A square representative of the background signal was also measured for each image. The 190 background signal was subtracted from the mean signal intensity of the cell to be used for 191

analysis. 192
To count vacuolation, the number of vacuoles in HFF and PPT1-deficient cells stained 193 with LAMP1 were counted using a mechanical hand counter. A minimum of 299 cells were 194 analyzed per conditioned group, across three replicates. Only vacuoles with a defined border 195 were counted. Images of LAMP1-positive vacuoles reported [32] were used as reference. For 196 autofluorescence analysis, a minimum of 121 cells were analyzed per conditioned group, across 197 two replicates, and background intensity was subtracted from the mean signal intensity to correct 198 for heightened background signal due to prolonged exposure time. based on the finding that 1 x 104 cells were confluent in a 96-well plate upon adherence. Because 208 the assay was used as a viability assay rather than a proliferation assay, cells were plated in 1% 209 serum containing DMEM, in order to measure the reduction in cell viability with 10,000 210 confluent cells as the baseline. To measure cell viability, cells were given a 48-hour incubation 211 period. After the incubation period, 0.5mg/ml of MTT reagent (10 µl) was added to each well, 212 and the plate was incubated for 4 hours to allow the reagent to be reduced. 100 µl of 213 solubilization solution was added to each well, and incubated overnight. Viability was 214 determined by the absorbance at 550nm minus the reference wavelength 690nm minus a plate 215 blank (as per manufacturer's protocol). Absorbance was measured using a Spectramax M5 plate 216 reader (Molecular Devices, California, USA). The experiment was repeated with a viability 217 measurement at 120 hours post-plating. To determine cell viability after hydrogen peroxide 218 (H 2 O 2 ) exposure, 1 x 104 cells were plated per well of 96-well plate, and given 24 hours to 219 adhere. Cells were then treated for 24 hours with 0, 25, 50, or 100 µM H2O2 in DMEM 220 supplemented with 10% FBS, 1% Glutamine, and 1% penicillin-streptomycin. The MTT 221 procedure was carried out as described to determine cell viability after 24  and MRC using the GFP filter. Autofluorescence was very low, barely detectable in wild type 249 primary fibroblast (n= 50), HFF (n = 56) and MRC-5 (n = 57) controls, but was marginally 250 visible in PPT1-deficient fibroblasts (n = 52) (data not shown). We quantify the fluorescence 251 levels in wild type HFF and MRC fibroblasts and PPT1-deficient fibroblasts to assess whether 252 the difference between normal and patient cells were statistically significant. Upon relative 253 fluorescence intensity (RFI) analysis, PPT1-deficient fibroblasts exhibited a 4.5-fold increase in 254 autofluorescence signal compared to controls. The One-Way ANOVA and post hoc Tukey HSD 255 analysis indicated that this RFI increase was statistically significant (p < 0.001). In contrast, RFI 256 between wild type control cell lines did not differ significantly (p = 0.996) (Fig 1). There was a detectable increase in autofluorescence signal in PPT1-deficient fibroblasts 260 compared to HFF and MRC-5 controls. Autofluorescence was increased >4.5-fold in PPT1-261 deficient fibroblasts compared to controls. Cells were stained with phalloidin-594 at 1:1250 to 262 use as a reference for focusing, and imaged using DAPI (ex 358nm / em 461nm) and GFP (ex 263 488nm / em 530nm) filters. HFF (n = 56), MRC-5 (n = 57), and PPT1 deficient (n =52) 264 fibroblasts were analyzed by measuring the relative fluorescent intensity using ImageJ. Error 265 bars display +SD. There were no significant differences in relative fluorescence intensity 266 (indicated by "n.s.") between the HFF and MRC-5 controls (p = 0.885). Fluorescent signal was 267 significantly higher in PPT1 deficient cells as compared to HFF (*, p < 0.001) and MRC-5 (*, p< 268 0.001) controls. 269

271
To investigate the possibility that increased autofluorescence observed in INCL 272 fibroblasts would lead to impaired cell viability, the MTT (C, N-diphenyl-N′-4,5-dimethyl 273 thiazol-2-yl tetrazolium bromide) proliferative assay was used to measure metabolic activity as 274 an indicator for cell viability [34]. PPT1-deficient fibroblast viability was reduced compared to 275 HFF and MRC-5 controls (Fig 2). Specifically, after 48 hours, PPT1-deficient fibroblast viability 276 (n = 7) was reduced significantly compared to that of HFF (n = 8) and MRC-5 (n = 8) controls (p 277 < 0.001). Significant differences were also observed between HFF and MRC-5 controls (p 278 <0.001). This experiment was repeated with an incubation time of 120 hours; the same relative 279 cell viability distribution was observed. PPT1-deficient cell viability (n = 5) was significantly 280 reduced compared to HFF (n = 7) and MRC-5 (n = 7) controls (p < 0.001). Significant 281 differences were also observed between control cell lines (p < 0.001). have an observable effect on autofluorescence accumulation. 295 We examined autofluorescence deposit levels in wild type (WT) and INCL patient (PT) 296 fibroblasts exposed to either WT and PT conditioned media (grouped as WT+WT, WT+PT, 297 PT+WT, and PT+PT; see Methods for details). We posited that the autofluorescence pathology 298 may be attenuated in the presence of functional PPT1 enzymes secreted in the media. Imaging 299 and quantitation of autofluorescence deposits was conducted using fluorescence microscopy at 300 excitation 488nm / emission 530nm (GFP) wavelengths (Fig 3). 301 As expected, we observed a three-fold increase in autofluorescence signal intensity (p < 302 0.001) between the levels of autofluorescence in the WT+WT (group 1) and PT+PT (group 4) 303 treatment groups (Fig 3B). Interestingly, PPT1-deficient fibroblasts exposed to wild type 304 conditioned media displayed significant reduction in autofluorescence as compared to PPT1-305 deficient fibroblasts incubated in PT-conditioned media (group 3 vs. 4). Signal intensity was 306 decreased to nearly half that of PT+PT cells. Nevertheless, PT+WT cells (group 3) still 307 exhibited 1.63 times greater autofluorescence level than the WT+WT (group 1) indicating that 308 secreted functional enzyme in the media is insufficient to completely restore autofluorescence 309 pathology to normal levels. Additionally, no statistically difference in intensity was observed in 310 WT cells grown in either WT-or PT-conditioned media indicating that endogenous functional 311 PPT1 enzyme were sufficient to overcome potential toxic effects secreted in PT-conditioned 312 media. Autofluorescence is higher in PPT1-deficient cells grown in either condition 3 or 4 as compared 317 to WT cells grown in either WT or PT-conditioned media (condition 1 or 2). Cells were stained 318 with DAPI at 1:1000 to use as a reference for locating and focusing on cells. Cells were then 319 imaged using the GFP (ex 488nm / em 530nm) filter. 320 Figure 3B. Quantitative analysis of autofluorescence storage material in four conditioned 321 media groups. 322 RFI was measured using ImageJ (n = 2 replicates per group). Significant differences (*, p < 323 0.001) were found between all conditioned groups, with the exception of group 1 vs 2, and group 324 2 vs 3, which are labeled n.s (not significant). Error bars indicate +/-SD. 325 326

PPT1-deficient fibroblasts displayed abnormal lysosomal distribution and elevated 327 numbers of lysosomal structures 328
We next investigated whether autofluorescence storage material was spatially consistent with 329 LAMP1-positive lysosomal structure; and whether increase autofluorescence correlated with 330 abnormal distribution of the lysosomes in PPT1-deficient fibroblasts. 331 Fluorescence microscopy was performed on primary wild type and PPT1-deficient 332 fibroblasts, and established HFF and MRC-5 fibroblasts revealed that patient fibroblasts 333 exhibited a higher level of lysosomal network as demonstrated by increased LAMP1 staining 334 intensity (Fig 4A). Normal fibroblasts displayed relatively sparse distribution of lysosomes 335 throughout the cell, with a slightly higher concentration of LAMP1-positive lysosomes in the 336 perinuclear region. In contrast, PPT1-deficient fibroblasts exhibited LAMP1-positive lysosomes 337 densely packed throughout the cell body (Fig 4). MFI was compared by one-way ANOVA (p < 338 0.001) and post-hoc Tukey HSD analysis, which showed that LAMP1-positive signal was 339 significantly greater in PPT1-deficient fibroblasts (n = 118) as compared to HFF (n = 114) and 340 MRC-5 (n = 94) controls (p < 0.001) (Fig 4B). Furthermore, a direct examination of LAMP1-341 positive lysosomal distribution in wild type and PPT1-deficient fibroblasts revealed a detectable 342 difference in fluorescent signal intensity between early passage (P3) PPT1-deficient fibroblasts 343 and wild type fibroblasts (Fig 5A). Analysis showed a statistically significant 1.3-fold increase 344 in LAMP1 signal intensity in PT cells (p < 0.01) (Fig 5B). The average LAMP1 signal intensity for the total area occupied by lysosomes was measured 350 using ImageJ (n = 94 -118 cells per group). Significant differences were found between controls 351 (*, p < 0.001). PPT1 patient cells had significantly higher signal compared to both controls (*, p 352 < 0.001). Error bars display +SD.

fibroblasts. 361
Average LAMP1 signal intensity was measured using ImageJ (n = 40 -60 cells for both groups). We next assess whether the observed abnormal lysosomal pathology in PPT1-deficient 368 fibroblasts can be lessened in the presence of wild type PPT1 enzyme in conditioned media using 369 LAMP1 antibody staining on all four conditioned groups 1-4 ( Fig 6A). LAMP1 fluorescence 370 intensity was statistically significant (p < 0.001) between all conditioned groups ( Fig 6B). 371

PT+PT cells (group 4) exhibited a two-fold increase in LAMP1 signal when compared to the 372
WT+WT control (group 1). In contrast, a significant reduction in LAMP1 signal was observed in 373 PT cells grown in WT-conditioned media (group 3) compared to PT+PT cells (group 4), but had 374 a 1.4-fold increase in intensity compared to WT+WT control (group 1). Relative to WT+WT 375 cells, WT cells conditioned with PT media (group 2) were found to have a 1.2-fold increase in 376 LAMP1 signal (Fig 6B). Relative fluorescence intensity (RFI) was measured using ImageJ (n = 3 replicates per group). 384 Significant differences (*, p < 0.001) were found between all conditioned groups. Error bars 385 indicate +/-SD. control cells observed (Fig 7). MRC-5 cells displayed normal highly branched interconnected 395 tubules. HFF cells had elongated tubules, but also exhibited normal branching (Fig 7 A&B). In 396 contrast, PPT1-deficient cells displayed a substantial decrease in mitochondrial tubule branching, 397 and the mitochondrial network instead consisted predominantly of non-tubular spherical punctate 398 structures (Fig 7C). shown to be decreased due to pharmacologically-induced mitochondrial dysfunction [27]. We 413 analyzed the expression pattern of a closely related lysosomal protease, cathepsin D, in PPT1-414 deficient fibroblasts because it has been implicated in the initiation of mitochondrial apoptosis 415 [35]. No differences were observed in the relative cathepsin D-positive signal density between 416 PPT1-deficient and HFF control fibroblasts (Fig 8). In both cell lines, cathepsin D-positive signal 417 was observed throughout the body of the cell, and partially within the extending membrane 418 processes. Although overlap between cathepsin D and LAMP1 signal was observed, cathepsin D 419 was not exclusively colocalized to LAMP1. As observed previously, PPT1-deficient fibroblasts 420 displayed a substantial increase in LAMP1 signal as compared to wild type HFF cells (Fig 8). 421 Large LAMP1-positive vacuoles have been shown to form due to ROS produced following 422 mitochondrial damage [27]. We used vacuole formation as an indicator for lysosomal 423 impairment, possibly brought on by mitochondrial dysfunction. The occurrence of large vacuoles 424 formed within LAMP1 stained cells was also increased in PPT1-deficient cells (Fig 8- The abnormalities found in the mitochondrial network were suggestive of mitochondrial 443 dysfunction [36], which is known to lead to increased ROS production [30]. We then tested 444 whether PPT1-deficient cells would be more susceptible to cell death induced by exogenous 445 ROS, as expected if pre-existing endogenous ROS were present. H2O2 is a ROS with biological 446 significance [37], and treatment with exogenous H2O2 is a well-established assay known to 447 induce apoptosis in a dose-dependent manner [38,39]. HFF and PPT1-deficient cells were 448 treated with increasing concentrations of 0 to 100 micromolar H2O2 for 24 hours in order to 449 examine susceptibility to oxidative damage by ROS (n = 5 wells per cell line per dose treatment). 450

Control cell viability declined in a dose-dependent manner with increasing H2O2 concentrations. 451
In contrast, PPT1-deficient cell viability was mostly depleted at all tested H2O2 concentration. A 452 univariate ANOVA revealed a significant group x dose effect (p < 0.001); however, group and 453 dose effects individually were not found to be significant (p = 0.071 and 0.054, respectively) (Fig  454   9A). We also measured endogenous ROS levels in our four conditioned groups 1-4 to 467 ascertain whether the presence of PPT1 in conditioned media had a positive influence on the 468 patient cell's susceptibility to H2O2 induced cell death. Results indicated while significantly 469 elevated (p < 0.01) relative luminescence units, an indicator for ROS, were detected between 470 both PPT1-deficient groups 3 and 4 as compared to wild type groups 1 and 2, there were not 471 significant difference in the levels of reactive oxygen species whether PPT1 patient cells were 472 grown in wild type (group 3) or PPT1 conditioned media (group 4) ( Fig 9B).  Fig S1). Similarly, the polymerized actin assembly also 487 appeared normal in PPT1-deficient cells (Supplementary Fig S2). were morphologically normal in PPT1-deficient cells indicating that the INCL pathology is 507 discreet and specific to abnormal lysosomal and mitochondrial networks. 508 Although GRODs are typically detected by electron microscopy [9,20,23], we report the 509 detection of autofluorescence storage material in PPT1 deficient fibroblasts using standard 510 fluorescence microscopy -a method similarly used in PPT1-deficient lymphocytes [21], and in 511 brain sections of INCL mice [13]. The presence of autofluorescence storage material also has 512 been reported in TPP1 and CLN3-deficient neural progenitor cells of late-infantile NCL and 513 juvenile NCL [40]. We confirmed the intralysosomal location of autofluorescence storage 514 material in PPT1-deficient patient cells by the co-localization of LAMP1 and autofluorescence 515 signals. This increased autofluorescence accumulation correlates with significantly reduced 516 PPT1-deficient patient cell viability as compared to either fibroblast control cell lines. However, 517 we observed that there are marked differences in cell viability between the MRC-5 and HFF 518 control fibroblasts, most likely due to specificity and robustness of each control fibroblast cell 519 lines. It should also be noted that the PPT1-deficient cells are untransformed and thus are not as 520 robust as either HFF or MRC-5 which may also impact cell viability. It can also be argued that 521 the observed difference reflects levels of metabolic activity rather than direct cell viability. 522 Because the MTT assay uses cell metabolism as an indicator for viability [33], substantial 523 metabolic differences could produce findings which may or may not accurately reflect viability. 524 If this was the case, we believe that decreased metabolic activity reflects a compromised 525 cytosolic state which would eventually lead to lowered cell viability. 526 Increased lysosomal staining intensity in PPT1-deficient fibroblasts was first reported using the 527 lysosomal marker LysoTracker suggesting an altered pattern of mitochondrial network [26]. 528 Consistent with these findings, we also observed increased staining intensity. Additionally, using 529 LAMP1, we observed a significant increase -three to four folds -in the number of lysosomal 530 structures, as well as dense distribution and localization of the lysosomes beyond the perinuclear 531 region. This indicates that substantially more lysosomal compartments were present in the PPT1-532 deficient fibroblasts, not just an increase in intensity due to abnormal accumulation and 533 distribution. It has been reported that PPT1 deficiency is closely linked with ER stress and 534 subsequent activation of the ER UPR [10,18]. PERK is known to play a key role in the 535 activation of the UPR [17], and the transcription factors TFEB and TFE3 have recently been 536 shown to activate lysosome biogenesis in a PERK-dependent manner [41]. The significant 537 increase in lysosomal compartments observed in our study may then represent evidence of 538 increased lysosomal biogenesis due to activation of the UPR, which supports the role for the ER 539 in INCL pathology reported previously [10,17,18]. Lysosome biogenesis occurred in a PERK 540 dependent manner which mediates ROS production and activation of mitochondrial-mediated 541 apoptosis in response to ER stress [17]. shown to be ROS-dependent [16], we next sought to investigate whether evidence for ROS 550 existed in PPT1-deficient human fibroblasts. Although ROS has been detected using a PPT1 KO 551 mouse model [10], ROS has not yet been reported in human cells or non-neuronal cell types. Our 552 results indicate that PPT1-deficient human fibroblasts exhibit a heightened susceptibility to cell 553 death induced by exogenous ROS. This is highly suggestive that elevated pre-existing 554 endogenous ROS are already present in PPT1-deficient patient cells. 555 To determine whether mitochondrial dysfunction could further impair lysosomal function, we 556 assessed the intracellular distribution of cathepsin D and the presence of large vacuole formation 557 from LAMP1-positive cells. Interestingly, cathepsin D was found abundantly throughout both 558 PPT1-deficient and control cells. By qualitative analysis, there were no differences in the spatial 559 distribution or expression of cathepsin D. Previously work has shown that cathepsin D is 560 involved in early stages of the mitochondrial-mediated apoptotic cascade [42]. Cathepsin D-561 deficiency has also been shown to lead to the accumulation of autofluorescence storage material 562 and progressive cell death, characteristic of the NCLs in general [42,43]. We find no correlation 563 with cathepsin D density and distribution and mitochondrial dysfunction. Since vacuolization is 564 ROS independent and has no morphological effects on the mitochondrial network [32], the 565 vacuolization observed in PPT1-deficient fibroblasts may represent a direct effect of 566 autofluorescence storage material accumulation on lysosome function rather than a complex 567 interaction with the mitochondria. 568 Finally, our work indicates that the loss of PPT1 enzymatic activity can be somewhat 569 mitigated with the introduction of wild type PPT1 enzyme in the cytosol. This method was first 570 reported as a potential enzyme replacement therapy for the lysosomal storage disorder 571 Mucopolysaccharidosis IVA [31] and may yet be a similarly viable avenue for INCL. Using this 572 paradigm, we ask whether the media collected from wild type cultures and PPT1-deficient 573 cultures have a positive or negative effect on wild type and PPT1-deficient patient fibroblasts. 574 The introduction of a functional enzyme secreted from wild type conditioned media may restore 575 normal enzyme activity by the observable reduction in autofluorescence storage materials. 576 Alternatively, patient conditioned media when added to wild-type cultures, may provoke an 577 abnormal phenotype due to the presence of secreted toxic factors. Our data indicates that the 578 patient cells benefitted from growing in the presence of wild type conditioned media: there are 579 dramatic reductions in autofluorescence accumulation and LAMP1 positive lysosomes as 580 compared to patient cells grown in their own conditioned media. Although cellular pathology 581 was partially mitigated, restoration was not at the wild type level. Levels of reactive oxygen 582 species were at comparably high levels whether PPT1-deficient fibroblast cells were grown in 583 wild type or PPT1-deficient conditioned media. These results indicate that complete rescue most 584 likely requires constitutive intracellular expression of PPT1 via a gene therapy vector or the 585 direct introduction of the enzyme to the brain or spinal cord. Currently, enzyme replacement 586 therapies for CLN diseases are invasive: intrathecal and intravenous administration of PPT1 in 587 the Ppt1-mouse model spinal cord, and the intracerebroventricularly administration of TPP1 588 enzyme for the treatment of NCL type 2 [1,44,45]. Our study indicates that normal PPT1 589 enzyme can be internalized by PPT1-deficient cells and be taken up by the lysosomes to repress 590 autofluorescence accumulation and abnormal lysosomal morphology. This paradigm has clinical 591 significance in that a partial cellular recovery may be possible using this passive method. 592 593