Selective Deletion of Heparan Sulfotransferase Enzyme, Ndst1, in Donor Endothelial and Myeloid Precursor Cells Significantly Decreases Acute Allograft Rejection

Early damage to transplanted organs initiates excess inflammation that can cause ongoing injury, a leading cause for late graft loss. The endothelial glycocalyx modulates immune reactions and chemokine-mediated haptotaxis, potentially driving graft loss. In prior work, conditional deficiency of the glycocalyx-modifying enzyme N-deacetylase-N-sulfotransferase-1 (Ndst1f/f TekCre+) reduced aortic allograft inflammation. Here we investigated modification of heparan sulfate (HS) and chemokine interactions in whole-organ renal allografts. Conditional donor allograft Ndst1 deficiency (Ndst1−/−; C57Bl/6 background) was compared to systemic treatment with M-T7, a broad-spectrum chemokine-glycosaminoglycan (GAG) inhibitor. Early rejection was significantly reduced in Ndst1−/− kidneys engrafted into wildtype BALB/c mice (Ndst1+/+) and comparable to M-T7 treatment in C57Bl/6 allografts (P < 0.0081). M-T7 lost activity in Ndst1−/− allografts, while M-T7 point mutants with modified GAG-chemokine binding displayed a range of anti-rejection activity. CD3+ T cells (P < 0.0001), HS (P < 0.005) and CXC chemokine staining (P < 0.012), gene expression in NFκB and JAK/STAT pathways, and HS and CS disaccharide content were significantly altered with reduced rejection. Transplant of donor allografts with conditional Ndst1 deficiency exhibit significantly reduced acute rejection, comparable to systemic chemokine-GAG inhibition. Modified disaccharides in engrafted organs correlate with reduced rejection. Altered disaccharides in engrafted organs provide markers for rejection with potential to guide new therapeutic approaches in allograft rejection.

In view of the fact that the graft recipients have normal Ndst1 expression (Ndst1 +/+ ), these findings indicate that deficiency of Ndst1 enzyme, specifically in the donor organ endothelial glycocalyx, reduced early allograft and vascular inflammation and rejection.

M-T7 treatment significantly reduces histopathological markers for renal allograft rejection.
Chemokines interact with both GAGs and cell receptors and are reported to alter transplant rejection, with potential effect on donor organ GAG and chemokine interactions during immune responses. Therefore, treatment  with the broad-spectrum chemokine-GAG inhibitor M-T7 was examined in WT renal allograft transplants for comparison to Ndst1 −/− donor allografts. Independent, blinded pathological analysis demonstrated significant reductions in histological markers of early allograft rejection at 10 days follow up with M-T7 treatment (10 daily doses, 100 ng/gm body weight) (Fig. 1B,D-J) in WT donor kidney transplants when compared to WT donor allografts treated with saline control (Fig. 1A). Changes produced by M-T7 were comparable to those seen with saline treatment of Ndst1 −/− donor allografts implanted into BALB/c recipient mice. M-T7 reduced overall pathology scores for early rejection ( Fig. 1D; P < 0.009) with independent reductions in cell infiltrates ( Fig. 1E; P < 0.001), vasculitis ( Fig. 1F; P < 0.036), glomerulitis ( Fig. 1G; P < 0.0001), peritubular capillaritis ( Fig. 1H; P < 0.013), and tubulitis ( Fig. 1I; P < 0.0001). M-T7 did not reduce the score for mesangial matrix ( Fig. 1J; P = 0.241), although treatment did indicate a trend toward reduction.

M-T7 and M-T7 point mutant treatment have variable efficacy in WT and Ndst1 −/− allografts.
M-T7-mediated reductions in markers of inflammation and early rejection were lost in Ndst1 −/− donor transplants (Fig. 2, N = 32 mice). This is consistent with interference with the known M-T7 mediated inhibition of chemokine to GAG binding. However, the independent beneficial effects of both Ndst1 deficiency and M-T7 treatment on reducing inflammation and rejection in renal allografts were lost in combination (i.e., M-T7 treatment in Ndst1 −/− donor organs) when compared to WT allografts ( Fig. 2B-H).
To further define interactions between M-T7, HS-GAG and chemokines, three M-T7 point mutations, with previously characterized variations in chemokine and GAG binding 29 ( Fig. 2A), were assessed for altered effects on rejection in WT C57Bl/6 and also Ndst1 −/− renal allograft transplant into BALB/c mice. Treatment with the M-T7 point mutations, F 137 D, R 171 E and E 209 I displayed differing inhibitory activities after transplant of C57Bl/6 WT donor kidneys into Balb/C mice ( Fig. 2B-H). R 171 E (P < 0.001) and E 209 I (P < 0.01) retained significant inhibitory function in donor WT renal allografts, whereas F 137 D (P = 0.1774) no longer blocked early rejection ( Fig. 2B-H).
R 171 E had no inhibitory activity in Ndst1 −/− donor transplants, as was seen for M-T7 (Fig. 2). R 171 E therefore had minimal differences from the native M-T7 inhibitory activity in this model. Conversely, E 209 I retained inhibitory activity in both WT C57Bl/6 and Ndst1 −/− donor renal allograft implants at 10 days follow up suggesting that E 209 I-mediated blockade of early signs of rejection is independent of Ndst1 or HS-GAG mediated chemokine interactions. F 137 D was inactive in both WT and in Ndst1 −/− allografts. In prior work, AlphaScreen assays for R 171 E and E 209 I demonstrated reduced binding to the chemokine RANTES when compared to M-T7 and F 137 D 29 . While all three tested point mutations had reduced RANTES binding in vitro, this effect was reduced for R 171 E and E 209 I. E 209 I had the smallest change in binding in the presence of heparin, suggesting that E 209 I may be less affected by heparin interaction with the chemokine-RANTES binding. Both R 171 E and E 209 I retained inhibitory activity for PMA-activated THP-1 cell migration in vitro, while F 137 D did not. There were also variations in independent histopathology findings for each mutant. F 137 D had an increase in vasculitis score (Fig. 2D) in Ndst1 deficient mice while E 209 I had a greater reduction in peritubular capillaritis in Ndst1 −/− allografts (Fig. 2F). Reduced HS and chemokine immunoreactivity is associated with reduced rejection. Transplanted sections were examined using immunohistochemical staining for HS and chemokines (Fig. 4). Glomerular HS staining in saline-treated Ndst1 −/− allografts was reduced when compared to saline treated WT transplants (P = 0.005; Fig  Significantly altered expression was detected for a subset of genes in signaling pathways as detected by qPCR analysis. M-T7-treated C57Bl/6 WT and Ndst1 −/− donor tissue (both showing reduced rejection) was compared to saline-treated C57Bl/6 WT donors (Fig. 5A). Changes due to treatment with M-T7 and F 137 D mutant (no reduction in rejection) versus saline-treated C57Bl/6 WT donors was also compared (Fig. 5C). Among the detected gene expression changes, Interleukin 4 (IL-4) was significantly decreased for both Ndst1 −/− and M-T7 treated WT grafts at 10 days follow up (Fig. 5A,B) versus saline-treated WT allografts. Heat shock transcription factor 1 (HSF1), Peroxisome proliferator-activated receptor gamma (PPARG), Telomerase reverse transcriptase (TERT), and WNT1 inducible signaling pathway protein 1 (WISP1) were significantly down-regulated in Ndst1 −/− grafts, but not M-T7 treated WT allografts. MDM2, CSF2, FOXA2, and TNF were significantly increased for M-T7 treated grafts, but not Ndst1 −/− grafts. Whereas Nitric oxide synthase 2 (NOS2), TRAF family member-associated NFκB activator (TANK), Early growth response 1 (EGR1), Fibronectin 1 (FN1), CC chemokine CCL20, Heat shock protein 90AA2 (HSP90AA2), IGFBP3, Selectin E (SELE) were decreased in M-T7 treated WT allografts (Fig. 5A). Specific gene expression changes were all within the NFκB and JAK/STAT pathways, but with changes primarily selective for either Ndst1-deficient grafts or M-T7 treatment in WT grafts. In the NFκB pathway, CCL20 was reduced by M-T7. In the JAK/STAT pathway, Interleukin-4 (IL-4) was significantly reduced in Ndst1 −/− grafts with saline treatment or in WT allografts with M-T7 treatment. NOS2 was also significantly reduced with M-T7 treatments in WT allografts. Although Murine double minute 2 (MDM2), a p53 regulator, was markedly increased by M-T7, this gene was also increased with F 137 D ( Fig. 5C) treatment, which does not reduce rejection, suggesting a poor correlation with reduced rejection.

Reduced early rejection is associated with modified macrophage and T cell invasion.
In summary, a series of genes in inflammatory signaling pathways demonstrated altered expression in grafts with reduced rejection. Significantly reduced IL-4 gene expression was detected for both Ndst1 −/− allografts and for M-T7 treated WT allografts (Fig. 5B). Significant changes for other genes differed in Ndst1 −/− grafts when compared to M-T7 treatment in WT grafts, suggesting differing targets.

Altered HS and CS disaccharide content is detected in renal allografts with reduced rejection.
Altered GAG content and metabolism was also examined for correlations with graft rejection. HS and CS disaccharide content and sulfation were measured in isolates from Ndst1 −/− allografts and from M-T7 or saline-treated WT allografts. Kidney samples vary in weight and thus disaccharides were normalized to total HS or CS Ndst1 −/− kidneys and M-T7 treated WT kidneys additionally had specific changes in percentage weight HS disaccharide, when compared to saline treated WT kidneys (Fig. 7) 35 . The percent weight (µg) of D0S6 was significantly increased in saline treated Ndst1 −/− (P < 0.022) and in M-T7 treated renal grafts (P < 0.006) when compared to saline treated WT grafts ( Fig. 7A; ANOVA P < 0.0143). Increased D2S6 was also detected in M-T7 treated WT (P < 0.013), but with a non-significant, borderline increase in saline treated Ndst1 −/− grafts (P = 0.168) (ANOVA P < 0.0414; Fig. 7B). D2A0 was borderline reduced (  compared to M-T7 treatment in WT grafts). As the method for labeling disaccharides can be complex, an explanatory diagram for HS disaccharide labeling is provided in Supplementary Fig. S1.
Significant changes were also observed in CS percent weights or CS disaccharides, although Ndst1 is reportedly selective for HS modification (Fig. 8). D0a4 (Fig. 8A) and D2a4 (Fig. 8B) CS disaccharides were reduced, D0a4 for both Ndst1 −/− transplants and M-T7 treated WT transplants (ANOVA P < 0.0043) and borderline for D2a4 (ANOVA P = 0.1390). Total CS content (Fig. 8I) was significantly reduced in M-T7 treated renal transplants (P < 0.019), but not in Ndst1 −/− transplants (ANOVA P = 0.078). Supplemental Figs S2 and S3 provide the HS and CS disaccharide data measurements using the same Y axis scale to allow for comparison of overall changes in content.
Combined changes in measured individual HS and CS disaccharides were correlated with overall pathology rejection scores measured on the same histology sections in Ndst1 −/− or M-T7 treated WT kidneys by multiple linear regression analysis (for HS disaccharides R = 0.992, R 2 = 0.984, for CS disaccharides R = 0.974, R 2 = 0.949).

Discussion
Early and ongoing activation of inflammatory immune cell responses, also termed innate or acute cellular rejection, are reported to induce ongoing organ damage and to be a significant driving force for late chronic transplant vasculitis, rejection and graft loss [36][37][38][39][40][41][42][43][44] . Late organ damage is also known to be caused by recurrent antibody-mediated immune rejection. Both antibody-mediated rejection and inflammatory cell responses are reported to contribute approximately fifty percent to ongoing chronic rejection, graft damage and vasculopathy, occurring concomitantly in 25% or more of rejection episodes.
The result of this study on early rejection of renal allograft transplant demonstrates significant reductions in rejection after implant of donor kidneys deficient in Ndst1, the primary sulfotransferase HS-modifying enzyme (Fig. 1). Recipient BALB/c mice in this study have normal (wildtype) Ndst1 expression. Thus, the reduction in early rejection histopathology scores after transplant of Ndst1 −/− donor organs with saline treatment is unique to the Ndst1-deficient donor organ, as no other treatment was given. In prior work, we demonstrated reduced aortic allograft inflammation and vasculopathy at later follow up times (4 weeks) in Ndst1 −/− donor aortic grafts and after M-T7 treatment in WT aortic allografts 19 . M-T7 also reduced chronic rejection and improved outcome in renal grafts at long term follow up in mice (100 days) and rats (5 months), respectively 19,26 , but was not previously tested for effects on early or acute rejection. The aortic transplant model is considered a model for chronic transplant vasculopathy, more closely representative of chronic arterial inflammation and repair rather than antibody-mediated rejection 43,44 . Thus, Ndst1 deficiency in the donor organ alone reduces both late (i.e., chronic) vasculopathy in aortic allografts and early or acute rejection in renal allografts. Significant and comparable reductions in rejection were also seen after treatment with M-T7, a broad-spectrum chemokine modulating protein that interferes with chemokine-GAG binding (Fig. 1). The capacity to modify rejection by altering GAG composition in the donor allograft may have broad potential for new treatment approaches in transplantation through modifying the donor organ.
While prior work has demonstrated that chemokines have an important role in immune and inflammatory responses in transplants, the role of the endothelial glycocalyx in donor organs has been less extensively studied 1,19,[36][37][38][39][40][41][42][43]45 . The capacity of an isolated decrease in Ndst1 expression specifically in the donor organ to significantly reduce early rejection suggests a central role for donor organ HS GAG content in rejection. Further the Ndst1 deficiency is selective for endothelial cells and myeloid precursors and one might predict that changes in donor Ndst1-deficient organs are predominately due to endothelial deficiency rather than myeloid precursors, as observed for thioglycollate-induced peritonitis and allergic contact dermatitis 14 . Because all donor immune cells could not be removed prior to renal transplantation, we cannot exclude the possibility of hematopoetic microchimerism and resident suppression as an involved mechanism in our model [46][47][48] . Indeed, it is known that LysM (leukocyte)-specific deletion of Ndst1 can affect inflammatory responses 49 . Thus, there is a possibility that resident immune cells in the graft, which may be devoid of Ndst1, may play a role in the maintenance of graft integrity. Wang et al. performed control experiments on bone marrow chimeras to note that the predominant inflammation-associated Ndst1 knockout effects in this specific strain of mice are almost exclusively due to knockout in the endothelium 14 . Nevertheless, this remains to be proven in a transplant model and will require further investigation in future studies.
A significant reduction was detected in HS staining and IL-8 CXC chemokine staining with Ndst1 −/− grafts suggesting that modified HS sulfation can interfere with chemokine-GAG gradient formation. However, M-T7 treatment, while reducing glomerular IL-8 CXC chemokine staining, did not significantly alter HS or CC MCP-1 chemokine staining ( Fig. 4; P = 0.063).
We assessed changes in HS and CS content in renal allografts, but unexpectedly found increases in select HS disaccharides suggesting that Ndst1 deficiency reduced HS sulfation in the endothelium, but led to an increase in disaccharides in the whole transplanted organ, potentially due to a response in the graft to the local endothelial changes. There were shared differences in disaccharide content and sulfation for Ndst1 −/− allografts and M-T7 treated WT allografts. HS disaccharide D0S6 was significantly increased both in Ndst1 −/− kidneys with saline treatment or in M-T7 treated WT kidneys. With the loss of M-T7 mediated suppression of rejection in treatment of Ndst1 −/− allografts, there was also a loss of the observed increase in D0S6 and D2S6 disaccharides. Select CS disaccharides, D0a4 and D2a4 were decreased for Ndst1 −/− grafts and for M-T7 treated WT grafts. Aside from the converging or similar changes in HS and CS disaccharides for Ndst1 −/− and for M-T7 treatment, our analyses also highlighted numerous diverging GAG composition changes. The increased 6-O sulfation for HS and 2-O sulfation for CS might suggest an inverse reaction in adjacent, non-deficient mouse tissues, e.g. around the endothelium, that may react to the Ndst1 deficiency. HS and CS disaccharide analysis was performed on whole Ndst1 −/− or M-T7 treated WT kidneys and not on endothelial cells alone. To assess whether an overall change in disaccharide content might correlate with the risk of rejection, MR analysis was performed. A correlation between disaccharide content and rejection scores was detected, again suggesting a correlation between overall GAG content and rejection. Increased 6-O heparan sulfation has been previously reported in renal transplant biopsies with increased chronic fibrosis and rejection 41 , thus indicating that altered D0S6 disaccharide content may represent one potential response to allograft rejection, whether protective or damaging 36 . Conversely, sites of low sulfation have been associated with potentially inflammatory endoglycoside heparanase degradation of the glycocalyx and thus increased 6-O HS sulfation observed in our study (Fig. 6B) may be protective 50 . Selective changes in disaccharide content may also be specific to individual cells as mast cell responses have demonstrated altered HS content with Ndst1 deficiency 51 . Studies reported by other groups do support a pro-inflammatory function for some GAGs. Heparanase treatment in a donor stem cell transplant model is reported to reduce rejection, improve cell survival 13,18 and reduce T H 2 responses and prevent diabetes in mice [36][37][38][39][40][41][42][43][44] . HS is up-regulated in transplant vasculopathy (TAV) in chronic rejection, as well as in ischemia reperfusion injury [36][37][38] . Antibodies to selected GAG species are reported to increase rejection 41 . Conversely, low molecular weight heparin infusion reduces scarring as well as transforming growth factor (TGF) and collagen expression after renal obstruction injury in mice. Further work will be required to determine the specific interactions between transplant alone and each treatment in WT and in Ndst1 deficient transplants. Due to the complexity of GAG synthesis and modification, we note that these findings do not demonstrate a direct cause and effect between altered disaccharide composition in donor organs with early rejection, but rather highlight a consistent correlation requiring further study. HS GAGs have multiple functions in addition to chemokine-mediated cell activation and migration and further in-depth analyses will be required to determine the exact mechanism(s) by which acute cell or antibody-mediated rejection is reduced with either Ndst1 deficiency or M-T7 treatment of allografts. We also report here a reduced efficacy for the M-T7 F 137 D point mutation in treating WT kidney grafts, suggesting that the beneficial effect of M-T7 in acute rejection with WT donors is specific to M-T7 mediated inhibition of chemokine binding to GAG alone and not through direct M-T7 interaction with GAG. The F 137 D mutant is predicted to have reduced chemokine interaction when compared to native M-T7 as the hydrophobic region of the structure that interacts with chemokines is disrupted. This supports the hypothesis that M-T7 reduces transplant rejection via inhibition of chemokines. In prior work F 137 D had reduced inhibition of chemokine binding when treated with heparin 29 . However, the finding here for F 137 D, differed from prior findings examining plaque growth in a mouse model of balloon angioplasty injury in hyperlipidemic ApoE null mice, where F 137 D retained some inhibitory activity, but did mirror the loss of inhibition for PMA activation on monocytes in vitro 29 . E 209 I was less affected by heparin competition in prior work in vitro 29 , and retained activity in Ndst1 −/− allografts in the current study. These observations suggest the possibility that E 209 I protein functions independent of GAG interactions. This difference in responsiveness for balloon angioplasty injury and solid organ transplant is not unexpected, as mechanisms underlying acute organ transplant rejection differ from those driving plaque growth after simple mechanical balloon angioplasty injury where there is no rejection.
Selective reduction in transplant organ CD3+ T cell invasion correlated with reduced rejection. The individual changes in modification of inflammatory response pathway genes support a key role for modification of innate and acquired immune responses in rejection, both via Ndst1 deficiency and with M-T7 treatment. The reduction in IL-4 gene expression by both approaches, M-T7 treatment in WT allografts and in saline treated Ndst1 −/− allografts, does suggest again shared or convergent regulatory pathways for Ndst1 −/− donor grafts and for M-T7 treatment in WT grafts. While the change in MDM2 is large and may represent a significant regulatory step, there is no comparable increase in Ndst1 −/− kidney and treatment with the inactive F 137 D point mutation also increased this gene, making this change of lesser interest. A transcriptome-wide analysis for genes that correlate with histopathological changes in rejection would be preferable, and will be approached in future analyses 34 .
These findings correlate the reduced rejection observed in saline treated Ndst1 −/− engrafted mice to altered HS and CS content, potentially via blockade of chemokine interactions. The observed reduction in overall rejection score is very similar to systemic treatment with M-T7, a broad-spectrum inhibitor of chemokine-GAG interactions. However, modulation of other GAG-dependent functions has yet to be examined in this model, and immune modulators other than chemokine-GAG interactions may play key functions. Further work will be necessary to examine other potential HS GAG interactions modified by Ndst1 deficiency or M-T7 treatment.

Conclusions
In conclusion, reducing endothelial and myeloid precursor cell Ndst1 expression, in donor allograft transplants alone, reduces acute renal allograft rejection, comparable to chemokine inhibition. Changes in donor organ HS disaccharide composition in transplanted organs has potential for diagnosis in detecting early rejection. These findings may also guide new donor-focused approaches to treating transplant organs designed to reduce acute inflammation and prevent chronic allograft damage, which is an unmet therapeutic need.

Materials and Methods
Animals. Three strains of mice were used in this study. C57BL6/J (stock #000664) and BALB/c (stock #000651) mice were obtained from JAX Laboratories (Bar Harbor, MN) or the University of Florida Animal Care Services breeding facility, which replenishes stocks from JAX Laboratories. Derivation and characterization of Ndst1 f/f TekCre + mice (Ndst1 −/− ; kindly provided by Dr. Jeffrey D. Esko, Glycobiology Research and Training Center, University of California, San Diego, CA) with floxed Ndst1 conditionally knocked out by Tek/ Tie2 endothelial tyrosine kinase promoter-driven Cre in the endothelial and sub-population leukocytes have been previously described 14 . All animals were housed in barrier conditions in vivaria of the University of Florida Animal Care Services. Mice were weaned at 3 weeks, maintained on a 12-hour light-dark cycle and were fed water and standard rodent chow ad libitum.

Surgical protocols -Kidney transplantation. All animal studies complied with University and National
Institutes of Health guidelines for the care and use of Laboratory animals and were approved by the University of Florida (UFL) and Arizona State University (ASU) Institutional Animal Care and Use committee (IACUC; UFL IACUC Protocol # 201604234_01; ASU IACUC Protocol #17-1549R). Renal allograft transplant was performed as previously described (Table 1; 6-10 mice with allograft transplant per donor organ genetic strain and treatment type; Total 80 mice). In brief, the donor kidney is placed in the left flank in the mouse and attached by end-to-side anastomosis between the donor suprarenal aortic cuff and the recipient aorta. Venous anastomosis between donor suprarenal inferior vena cava (IVC) and recipient IVC is performed in the same fashion and the bladder attached, as previously described 19 .
A series of donor renal allografts from either C57Bl/6 wild type (WT) or Ndst1 −/− were transplanted into BALB/c mice. Mice with WT or Ndst1 −/− donor allografts were treated with either saline control, M-T7 or individual mutated constructs (M-T7-His 6X , F 137 D, R 171 E, or E 209 I; 6-10 mice per donor organ strain and per treatment) ( Table 1) 29,34,52 . Donor renal allografts were transplanted into BALB/c mice after resection of both kidneys under general anesthetic. No other immune suppressants were given to the mice before or after transplantation, in order to examine isolated early effects of each condition alone on transplant rejection at 10 days follow up 19,[31][32][33] . Treatments were given daily by intraperitoneal (IP) injection at 100 ng/gm/day X 10 days per mouse for each individual protein treatment 19 . Mice were euthanized at 10 days follow up with Euthanyl (Virbac AH Inc., Fort Worth, TX) as previously described 19 . For tissue analyses, renal allografts were divided into 3 sections and each expressed and purified as previously described. In brief, M-T7 mutants were generated by mutagenic PCR using M-T7pFastBacDualeGFP as the template 112 . Mutant constructs and wild type M-T7 were transformed into DH10Bac bacteria (Invitrogen.Carlsbad, CA), and blue/white screened on LB + Kan + Tet + Gen + IPTG+ X-gal plates. Bacmids were purified and used to transfect Sf9 insect cells with Cellfectin II (Invitrogen. Carlsbad, California). Baculovirus supernatants were collected to infect insect cells and express the various M-T7 mutant proteins. M-T7 and each of the three mutant constructs were then purified by sequential column purification as previously described 19,29 . Histological and immunohistochemical analysis of acute rejection and scarring. Sections of transplanted organs were cut into three 1.5-2 mm equal length cross sections for histology, fixed, paraffin embedded, and cut into 4-5 µm sections (3 sections per transplant specimen, providing 9 sections per allograft). Histology sections were stained with Haematoxylin and Eosin (H & E), Masson's trichrome, and Periodic acid-Schiff (PAS) as previously described 19,[23][24][25][26][27][28][29] . All sections were analyzed for changes consistent with acute rejection and vasculitis 34 by pathologists blinded to the mouse donor allograft implant and to treatment with either saline or M-T7 using Banff diagnostic criteria (DW, WC, BC). Pathology was scored on a scale of 4 and overall pathology score was a summation of independent scores assessed by detection of cellular infiltrate, vasculitis, glomerulitis, peritubular capillaritis, tubulitis, and mesangial matrix.
RT-PCR array analysis of altered gene expression in renal allografts. One third of each transplanted kidney section was collected in RNAlater (Ambion, Austin, TX) and RNA was isolated using RNeasy Mini kit following the manufacturer's protocol (Qiagen, Valencia, CA). RNA was reverse transcribed to cDNA using Superscript VILO cDNA Synthesis kit (Invitrogen Corporation, Carlsbad, CA) and Real Time PCR carried out using SYBR Green Core Reagent kit and a 7300 RT-PCR system (Applied Biosystems, Austin, TX). Changes in gene expression were normalized to internal GAPDH control and subsequently to saline treated controls. Primers specific to inflammatory and apoptotic pathways were assayed and are listed in Supplemental Table 1.

Analysis of GAGs from saline treated Ndst1 −/− and WT kidneys, with and without M-T7. Total
HS-and CS-GAG content and percent weight disaccharide were measured in transplanted kidneys from WT mice, with and without M-T7 treatment, and in transplanted kidneys from Ndst1 −/− mice with saline or M-T7 treatment (N = 10, 3-4 mice per strain and treatment group). Researchers were blinded to samples (SA, PA). HS and CS GAG composition were quantified by HPLC 53 . Formalin-fixed, paraffin-embedded samples were extracted as previously described 54 . Samples were incubated in glass tubes in a heating block at 60 °C with 3 mL of xylene, heated for 15 minutes with 2 repeats, and then rinsed with 3 mL of 100% ethanol, followed by 96, 70, 50, and 30% ethanol in water and washed three times with 18 MΩ water. Samples were rehydrated in 1 mL solution of PBS for >30 minutes at RT. Rehydrated tissue was homogenized and defatted in acetone over 48 h with shaking at 4 °C. Samples were dried and suspended in 2 mL of 0.1 M Tris-HCl, pH 8.0, containing 2 mM CaCl 2 , and 1% Triton X-100. Pronase (0.8 mg/mL) was added to the suspension and tissue digested at 50 °C for 24 h with one 24 h repeat. Finally, pronase was inactivated by heating to 100 °C for 15 min. Buffer was then adjusted to 2 mM MgCl 2 , benzonase (100 mU) added, and samples incubated for 2 h at 37 °C. After inactivation of the enzyme (15 min, 100 °C), undigested tissue was removed by centrifugation for 1 h at 4000 g. Sample supernatant was applied to a DEAE-Sepharose -micro spin column (Harvard Apparatus), washed with ~10 column volumes (CVs) of loading buffer (~pH 8.0 Tris Buffer) twice, allowed to adhere, washed with ~20 CVs of loading buffer, followed by ~20 CVs of wash buffer (~pH 4.0 Acetate Buffer), and 3 CVs of water. Samples were further cleaned with ~15 CVs of 0.2 M NH 4 HCO 3 . To elute, ~20 CV of 2 M NH 4 HCO 3 was added to the column and the fractions were collected. Samples were then freeze-dried and dissolved in 100 µL water.
Samples were digested with Heparinases I-III (New England Biolabs) or Chondroitinase ABC (Sigma), producing disaccharides that were separated using SAX-HPLC coupled to post-column fluorescence labeling and detection (Agilent system using a 4.6 × 250 mm Waters Spherisorb analytical column with 5 μm particle size at 25 °C). Peak migration times and areas were calculated compared to known disaccharide standards. Representative chromatograms are provided in Supplemental Fig. 4. HPLC were run with two solvents, Solvent A: 2.5 mM sodium phosphate, pH 3.5 and Solvent B: 2.5 mM sodium phosphate, 1.2 M NaCl, pH 3.5 with gradated change from 97% A and 3% B to 100% B and 0% A (65 min, flow rate 1.0 mL/min). GAG detection was performed by post-column derivatization. Eluents were combined with a 1:1 mixture of 0.25 M NaOH and 1% (w/v) 2-cyanoacetamide pumped at a flow rate of 0.5 mL/min from a binary HPLC pump (SSI Scientific Systems, Inc) and heated to 130 °C in a 10 m reaction coil, before cooling and return into the Agilent's fluorescence detector (λ ex = 346 nm, λ em = 410 nm). Commercial standard disaccharides (Dextra Laboratories) were used for identification based on elution time, as well as for calibration.
Paraffin-containing formalin from each sample was assessed in parallel as a control. No GAG or disaccharide was detected in these controls. Following isolation the GAGs released with β-elimination (1% w/w sodium borohydride in 2N NaOH) were desalted with a PD-10 column (GE Healthcare), and freeze-dried before disaccharide composition analysis. Peak migration times and areas for disaccharides separated by HPLC were compared to known standards (See example in Supplemental Fig. S4).
Statistical analysis. Measured change in histopathology scores, incorporating tissue mononuclear cell count, percentage of positively stained cells, PCR array, HS staining, and tissue HS and CS disaccharide content were assessed for statistical significance using analysis of variance (ANOVA) with secondary Fisher's PLSD or Student's unpaired, two tailed T-test, as previously described 19,[23][24][25][26][27][28][29] . Multiple and simple regression analyses were performed to assess correlations between tissue disaccharide content with total histopathologic score for acute rejection 35 .