Alport syndrome is an inherited disorder caused by mutations to the COL4A3, COL4A4, and COL4A5 genes. It is marked by dysregulation of GBM proteins, including collagen IV and laminin, resulting in altered GBM morphology, proteinuria, and renal failure (Kashtan et al., 2001; Abrahamson et al., 2003; Hudson et al., 2003). Normally, laminin α₁β₁γ₁ (LM-111) is expressed only during the early stages of glomerular development. In Alport syndrome, the LM-111 is re-expressed in the adult GBM (Abrahamson & St John 1993; Abrahamson et al., 2003; Abrahamson et al., 2007), although the mechanisms governing this re-expression are unclear. It has been shown that sulfatides, a class of acidic glycosphingolipids, can initiate basement membrane assembly by binding directly to LM-111 (Li et al., 2005). This finding suggested that sulfatides within the plasma membrane may play a role in laminin deposition. Given the dysregulation of laminin expression in the Alport GBM, we sought to determine if the renal sulfatide levels would follow an abnormal temporal pattern in Alport syndrome.

Sulfatides (3-O-sulfogalactosylceramides) consist of a ceramide with a sulfated galactose head group. Sulfatides have been found in a variety of tissues, including kidney, and have various structures, often depending on tissue type (Kyogashima et al., 2006). Changes in the sulfatide structure may influence its presentation at the cell surface and the degree of clustering of receptor molecules (Pruett et al., 2008). Alterations in sulfatide metabolism have been suggested to play a role in a number of diseases, including renal cell carcinoma, hepatorenal fibrocystic disease, and protein overload nephropathy (Li et al., 2009; Zhang et al., 2009; Takahashi & Suzuki 2012).

To investigate sulfatides in the intact renal tissue during kidney development, we used Matrix Assisted Laser Desorption Ionization Imaging Mass Spectrometry (MALDI IMS) (Caprioli et al., 1997). This technology combines the high sensitivity and selectivity of mass spectrometry with the spatial analysis provided by traditional histology, offering visualization of the distribution of specific biomolecules in tissue (McDonnell & Heeren 2007; Norris & Caprioli 2013; Gessel et al., 2014). This technology has been previously applied to the studies of lipids in renal diseases (Herring et al., 2007; Murphy et al., 2009; Kaneko et al., 2011; Marsching et al., 2011; Ruh et al., 2013; Grove et al., 2014; Jones et al., 2014).

We compared the temporal and spatial distribution of sulfatides in renal sections from the wild-type (WT) mice and a mouse model of Alport syndrome. The levels of two sulfatides, SulfoHex-Cer(d18:2/24:0) and SulfoHex-Cer(d18:2/16:0), dramatically increased in the Alport compared to WT kidneys during development up to 28 days of age. These temporal changes mirrored the known abnormality of laminin assembly in Alport syndrome. The sulfatide species were localized to renal tubules, thus suggesting that this aberrant lipid pattern may be related to the development of tubulointerstitial fibrosis in Alport disease. To the best of our knowledge, this is the first report suggesting a role for specific lipid species in the pathogenesis of Alport syndrome.

Multiple sulfatide lipids were detected and most showed the same levels in WT and Alport mice at 2, 14 and 28 days of age (Table S1). However, we identified two signals at m/z 888.6210 and 776.5005 with abundances that were similar in both groups at 2 and 14 days but dramatically increased at 28 days in Alport but not in WT mice (Fig. 1, Fig. S1). The signal was co-localized in the renal cortex (Fig 1, Fig S1). Other lipids were also detected in the sample (see Table S1 for a list of these species), but their overall levels did not appear to change (data not shown).

A database search (http://lipidmaps.org/) for m/z 888.6210 (3 ppm error) indicated that the mass corresponded to three potential lipid structures. These included three sulfatides with varying chain lengths and saturations: (d16:1/26:1), (d18:1, 24:1), and (d18:2, 24:0). To identify which compound was present in our tissue samples, MALDI MS/MS experiments were performed directly from the tissue surface (Fig. 2). Several sulfatide-specific fragments were present in the spectrum including peaks at m/z 97 and 241 (Hsu & Turk 2004). In addition, peaks specific to a doubly unsaturated sphingoid base (18:2) were present including m/z 538, 520, and 392 identifying the molecule at as SulfoHex-Cer(d18:2/24:0) (Fig. 2 A). Similar analysis of the peak at m/z 776.5005 identified this signal as SulfoHex-Cer(d18:2/16:0) (Fig. 2 B). Importantly, the high resolution experiments allowed us to exclude peaks at m/z 888.5561 and m/z 888.5709 that were nominally isobaric with the C¹³ isotopes of phospholipids. These phospholipid species did not change in WT vs. Alport kidneys and had very different tissue localizations compared to sulfatide species of interest (Fig. 3).

To determine specific localization of the two sulfatides of interest within the kidney, MALDI IMS data was acquired at 15 μm resolution. Following IMS analysis, the MALDI matrix was removed, the tissue stained with PAS and IMS images correlated with histological tissue features. The signal of SulfoHex-Cer(d18:2/24:0) (m/z 888.6224, 1.8 ppm error) was localized to tubular regions of the kidney but not to glomeruli (Fig. 4, Fig. S3). The tubular localization was the same in 2-day and 28-day old mice (data not shown). It was not possible to distinguish between proximal and distal tubules, due to the thickness of the section and the damage from the MALDI IMS laser.

We have identified two renal tubular sulfatide species, SulfoHex-Cer(d18:2/24:0) and SulfoHex-Cer(d18:2/16:0), as potential players at the very early stages of tubulointerstitial pathology in Alport syndrome. No other sulfatides with a diene sphingoid base were detected in this experiment, although it is possible that others do exist at concentrations below the sensitivity of MALDI IMS. The primary structural difference between these dysregulated sulfatide species and those sulfatides which levels were not affected in Alport is the presence of a second double bond in the sphingoid base of the dysregulated sulfatides. Sulfatides with a 4E, 14Z-diene in the sphingoid base of the ceramide moiety have been previously detected in mouse and human tissues (Panganamala et al., 1969; Renkonen & Hirvisalo 1969; Marsching et al., 2011).,

MALDI IMS is less sensitive than LC MS/MS, due to the small amount of analyte ions desorbed in one ablated spot and due to ion suppression from the heterogenous biomolecules that exist in the sampled tissue area. However, it does provide information about analyte tissue localization, which is key in this study. Another limitation of MALDI IMS is that absolute quantitative analysis is a significant challenge, again due to ion suppression affects, uneven tissue morphology which affects ionization, and challenges associated with standardization of analyte signal from tissue sections

In this study, the signal intensities of SulfoHex-Cer(d18:2/24:0) and SulfoHex-Cer(d18:2/16:0) were clearly different in WT and Alport mice at 28 days. This is apparent in both the heat map image shown in Figure 1 and in the extracted mass spectrum which is an average of all mass spectra collected from tubular regions (Figure S1). Other lipids, including sulfatides were detected in the MALDI IMS experiments (Table S1) and their signal intensities were not clearly different in Alport vs. WT kidneys (data not shown). Overall, our data support the conclusion that the expression of these specific sulfatide species is clearly altered between the groups.

Changes to sphingoid base saturation can confer bioactive properties to related lipid species. For example, sphingadienes, which are precursors of SulfoHex-Cer(d18:2/24:0) and SulfoHex-Cer(d18:2/16:0), have been shown to inhibit Akt-dependent and Wnt signaling in colon cancer cells (Fyrst et al., 2009; Kumar et al., 2012). In both cases, sphingosines containing single double bond did not show inhibitory effect. In another study, ceramides with an additional trans-double bond in the sphingoid backbone induced apoptosis in breast cancer cells (Struckhoff et al., 2004). In the current study, the diene sphingoid base was present only in the two sulfatide species that were dysregulated in Alport mice and not in any other identified sulfatide species. Overall, our results suggest that the second double bond is required for dysregulation of sulfatide levels in Alport mice.

Interestingly, the increase in renal levels of SulfoHex-Cer(d18:2/24:0) and SulfoHex-Cer(d18:2/16:0) at 28 days is similar to re-expression of LM-111 in the glomerular and tubular BM of Alport mice (Abrahamson & St John 1993; Abrahamson et al., 2003; Abrahamson et al., 2007), pointing to a potential relationship between these two phenomena. Although laminin expression was not measured in this study, the localization of SulfoHex-Cer(d18:2/24:0) and SulfoHex-Cer(d18:2/16:0) tubules would suggest that these sulfatides would interact with tubular rather than glomerular LM-111 as they specifically localize to the tubular regions of the kidney.

Although Alport disease is initiated in the glomerulus, later stages are characterized by progressive tubulointerstitial fibrosis in Alport mouse and in human kidneys (Kashtan & Kim 1992; Miner & Sanes 1996). Notably, tubulointerstitial fibrosis was observed in 79-day old Alport mice, while kidneys of 36-day old mice showed no clear signs of renal injury (Miner & Sanes 1996). Our detection of sulfatide dysregulation at 28 days, prior to appearance of histopathological lesions, suggests that these changes may be involved in very early stages of tubular injury. Because of the thickness of the tissue section and the damage from the MALDI laser, it was not possible to gain a more detailed view of the tissue histology from the sections that were imaged by MALDI IMS (e.g. localizing the signals to proximal vs. distal tubules or measuring the extend of tubular injury).

We speculate that one of the potential mechanisms of tubular injury involving sulfatides may be tubular reabsorption of profibrotic chemokines and cytokines secreted by stressed Alport podocytes, although this is not proven here (Ninichuk et al., 2005; Kruegel et al., 2013). Besides directly binding several chemokines, sulfatides can specifically affect monocyte infiltration into renal interstitium and trigger up-regulation of the chemokine co-receptor CXCR4 on leukocytes (Ogawa et al., 2004; Duchesneau et al., 2007; Takahashi & Suzuki 2012). Sulfatides can also bind to receptors on macrophages, thus triggering production of TGF-β1, a key cytokine in the pathogenesis of renal fibrosis (Popovic et al., 2007).