Linkage of a plasma zinc signature and impaired insulin receptor activation: Implications for the mechanism of type 2 diabetes mellitus

Type 2 diabetes mellitus (T2DM) is characterized by decreased plasma zinc levels and hyperzincuria, yet the underlying cause of these zinc disturbances is unknown. In this study, we compared postprandial plasma zinc levels in samples from T2DM and healthy control subjects to determine whether zinc is associated with a different set of proteins. We found that in T2DM a considerable amount of zinc remained following albumin/Ig depletion. A discrepancy in total amount of zinc in the remaining set of zinc-associated proteins identified and estimated by protein analysis as alpha-2 macroglobulin (A2M), and in T2DM alone some bacterial proteinases as well, indicated that the likely source of this discrepancy was from bacterial zinc proteinases trapped by A2M that obscured the high levels of these proteinases. Furthermore, an insulin receptor assay examined whether activated A2M (A2MFF) affected insulin receptor activation. The results showed a significant decrease in insulin receptor activation following repeated treatments with A2MFF but not after a single treatment with A2MFF. Our findings suggest that in T2DM, A2MFF likely arises from the trapping of zinc-dependent bacterial proteinases and impairs insulin receptor activation from a prolonged presence, which may result in a “receptor-protective” effect manifested as insulin resistance.

The aim of this study was to identify the source of decreased zinc and zinc loss and to 44 discover a mechanism for the loss of zinc observed in T2DM. This study compared postprandial 45 plasma zinc levels between healthy controls and T2DM subjects to determine possible 46 differences in zinc-associated proteins. This study has identified the likely source of zinc loss as 47 bacterial zinc proteinases that cause the formation of activated A2M, which is then presumably 48 cleared through the LRP1 receptor pathway. Additionally, we also found that in the continued 49 presence of activated A2M, activation of the insulin receptor is impaired and results in a reduced 50 insulin response. 51 52

Insulin Response Assay 103
The PathHunter ® Insulin Bioassay Kit was purchased from Discoverx/Eurofins (Fremont, CA, 104 US). This kit was used to measure the insulin response through chemiluminescent output in a 105 cell-based functional assay for insulin receptor isoform b. The insulin response was measured for 106 A2M at a concentration of 1.2 µg/ml with either A2MFF added once, at a concentration of 0.24 107 µg/µl, or A2MFF added every 30 minutes, at a concentration of 0.24 µg/µl, throughout the 3-108 hour incubation after the addition of insulin. Native A2M was purchased from Genway Biotech 109 (San Diego, CA US), and A2MFF was purchased from Sigma-Aldrich (St. Louis, MO US). 110 111

Statistical Analysis 112
Statistical differences in plasma zinc levels between experimental groups (Fig 1) were 113 determined by paired, two-tailed Student's t-test, and values of P < 0.05 were considered 114 significant. Statistical differences in the insulin receptor activation in response to insulin (Fig 2)  115 were assessed by paired, two-tailed Student's t-test two-way ANOVA followed by Dunnett's 116 multiple comparisons test. In both methods values of P < 0.05 were considered significant.

Comparison of Postprandial Zinc Levels 141
We measured postprandial plasma zinc levels in subjects with T2DM and healthy controls before 142 and after albumin/Ig depletion. We found zinc levels in all subjects before albumin/Ig depletion 143 were in the normal range of 60-130 µg/dL. Zinc levels after albumin/Ig depletion were nearly 144 zero in the controls and significantly above zero in the T2DM subjects. This is shown as the 145 difference between zinc levels between unprocessed plasma and depleted plasma with a known 146 quantity of added plasma to bring values above the detection threshold (Fig 1). In the controls 147 8 (samples 1-3), there was no significant difference between pre-and postdepletion plasma (paired 148 two-tailed t-test, P = 0.0572,). The T2DM group (samples 4-9), however, showed a significant 149 difference in pre-and postdepletion zinc levels (paired two tailed t-test, P = 0.0232). A 150 comparison of the controls vs. T2DM groups showed a significant difference between the groups 151 of 6.5 units (paired two-tailed t-test, P = 0.0280). A Gaussian distribution was assumed even 152 though the sample sizes were small. These findings suggest that there is a small fraction of 153 proteins with zinc affinity other than albumin/Ig in T2DM that are not present in healthy 154 controls. 155 156

Liquid Chromatography with Tandem Mass Spectrometry Based Protein Identification 157
To identify the zinc-associated proteins in the samples after albumin/Ig depletion, we analyzed 158 postprandial albumin/Ig-depleted plasma from each subject after filtration using Zn-IMAC and 159 liquid chromatography with tandem mass spectrometry (LC/MS/MS) for protein identification 160 and analysis. Of the 7 samples analyzed, 3 control subjects and 2 prediabetic patient showed zero 161 percent of bacterial proteinases, while 2 T2DM samples showed a minor amount of bacterial 162 proteinases. In the T2DM samples, the bacterial proteinases alone did not add up to the amount 163 of remaining zinc not associated with albumin/Ig. All samples showed that the majority protein 164 was alpha 2 macroglobulin (A2M) (  showed a large, significant decrease in insulin receptor activation following A2MFF treatment 183 every 30 minutes relative to a single treatment of A2MFF (paired two-tailed t-test, P = 0.0173). 184 There was no significant difference between the effect of A2M alone and A2M with a single 185 A2MFF treatment on insulin receptor activation (paired two-tailed t-test, P = 0.0615) (Fig 2). We 186 performed a two-way ANOVA analysis and observed a significant difference in insulin receptor 187 activation at two different concentrations (300 ng/mL, 100 ng/mL) between A2M with a single 188 treatment of A2MFF (two-way ANOVA, P = 0.9548, P = 0.9549), which showed no significant 189 difference compared with A2M alone, and A2M with multiple treatments of A2MFF (two-way 190 ANOVA, P = 0.0027, P = 0.0148), which showed in a significant difference compared to A2M 191 alone (Fig 2). These results demonstrate that insulin receptor activation is significantly reduced 192 in the presence of A2MFF, likely due to interference with the ability of insulin to access the 193 insulin receptor. 194 We also performed insulin receptor activation assays individually with A2M, A2MFF. 195 We found there was no significant difference in insulin receptor activation between A2M and a 196 single treatment with A2MFF (paired two-tailed t-test, P = 0.2514). Furthermore, in a separate 197 experiment, we used almost twice the concentration of A2MFF (0.44 µg/µl) compared to that in 198 a previous experiment, with an increased volume (15% of the reaction volume compared to 3% 199 of the reaction volume in the previous experiment), in the presence of A2M. We found a similar 200 reduction in insulin receptor activation following both single (paired two-tailed t-test , P = 201 0.0043) and multiple treatments with A2MFF (paired two-tailed t-test, P = 0.0016) when 202 compared with A2M alone. The reduction in insulin receptor activation in the large single 203 treatment group was prolonged and in the same range as that of the group with multiple 204 treatments of A2MFF (data not shown). 205 206

Discussion 207
A prominent characteristic of T2DM is decreased plasma zinc levels and hyperzincuria, for 208 which a cause is unknown [7]. We hypothesized that zinc loss is due to the association of zinc 209 11 with a different set of proteins that may be in a biologically unavailable form. To explore this 210 possibility and considering that a majority of zinc in plasma is associated with albumin/Ig [7,8], 211 we compared postprandial albumin/Ig-depleted plasma zinc levels between healthy controls and 212 T2DM subjects. Our results showed some amount of remaining zinc in albumin/Ig depleted 213 plasma in T2DM subjects but not in healthy controls (Fig 1). This finding indicated that a small 214 but significant portion of zinc is bound to proteins other than albumin/Ig in T2DM. 215 To identify the proteins associated with the remaining zinc in albumin/Ig depleted 216 postprandial plasma of T2DM subjects, we purified the depleted plasma on a zinc immobilized 217 metal affinity chromatography (Zn-IMAC) column and analyzed the proteins using LC/MS/MS. 218 The results showed that the majority of protein was A2M in both the T2DM subjects and 219 controls. This finding is consistent with similar observations in a previous study [8]. Some 220 bacterial proteinases were observed in 2 T2DM subjects but none in healthy controls or 1 221 prediabetic subject (Table 1). The total amount of bacterial zinc proteinases did not add up to the 222 level of remaining zinc. Prediabetic sample were not different from the control samples for 223 bacterial zinc proteinases, even though we detected some remnant zinc after albumin/Ig 224 depletion. A previous study showed increased A2M levels with a reduced trypsin binding 225 capacity in diabetes [6]. We inferred that these results, albeit indirectly, indicated that the 226 remaining zinc after albumin/Ig depletion likely came from activated A2M proteins that have 227 trapped bacterial zinc proteinases, since A2M is a pan-protease inhibitor that traps free 228 proteinases, resulting in A2MFF. In the prediabetic case, free bacterial proteinases may not be 229 detected because they may be obscured when captured by A2M, without exceeding the capture 230 threshold for levels of activated A2M [9]. 231 Studies have shown that proteinase-activated A2M (A2MFF) binds numerous receptors, 232 including, LRP1, growth factors and cytokines [10]. We hypothesized that A2MFF may also 233 bind the insulin receptor. We tested this possibility with an insulin receptor activation assay. We 234 measured insulin receptor activation following treatment with non-activated A2M with either 235 single or multiple treatments with activated A2M (A2MFF). Since activated A2M is rapidly 236 cleared away by the LRP1 receptor in approximately 2-4 minutes [11,12], multiple treatments 237 with activated A2M were used to replace activated A2M that would be cleared away during the 238 long incubation, thereby providing a continuous presence of activated A2M and revealing its 239 effects. 240 We observed a significant decrease in insulin receptor activation after multiple treatments 241 with A2MFF (0.24 µg/µl) compared to a single treatment with A2MFF (0.24 µg/µl) (Fig 2). We