J Alzheimers Dis. 2014;42(4):1279-94. doi: 10.3233/JAD-140672.

Fibroblast Aggregation Rate Converges with Validated Peripheral Biomarkers for Alzheimer’s Disease

Florin V. Chirila*, Tapan K. Khan, and Daniel L. Alkon

Blanchette Rockefeller Neurosciences Institute, Morgantown WV 26506

*Corresponding Author
Blanchette Rockefeller Neurosciences Institute. 8 Medical Center Drive, Morgantown, WV 26506, USA.
Tel.: +1 304-293-8404; Fax: +1 304-293-3675
E-mail address: fchirila@brni.org (Florin V. Chirila)



The inaccuracy of the diagnosis for Alzheimer’s disease (AD) has made its therapeutic intervention difficult, particularly early enough to prevent significant neurodegeneration and cognitive dysfunction. Here, we describe a novel, highly accurate peripheral diagnostic for AD patients based on quantitatively measured aggregation rate of human skin fibroblasts. The elevated aggregation rate with increasing cell density in AD cases is the basis of this new biomarker. The new biomarker was successfully cross-validated with two more mature assays, AD-Index, based on the imbalances of ERK1/2, and Morphology, based on network dynamics, and showed 92% overlap. A significant number of cases tested with this new biomarker were freshly obtained (n = 29), and 82% of the cases are hyper-validated cases, i.e., autopsy and/or genetically confirmed AD or non-Alzheimer’s disease demented patients (Non-ADD) and non-demented age-matched controls. Furthermore, we show that by using a simple majority rule, i.e., two out of the three assays have the same outcome, we significantly increase the agreement with clinical AD diagnosis (100%). Based on the high accuracy of this strategy, the biomarker profile appears to accurately identify AD patients for therapeutic intervention.

KEYWORDS: Aggregation rate; Alzheimer’s disease; cross-validation; majority rule; skin fibroblasts

PMID: 25024330



There are four ideas emphasized in this paper. The first idea is the new peripheral bio-marker for Alzheimer’s disease which quantifies the Aggregation Rate for human skin fibroblasts. This bio-marker is a refinement of the previously reported measure of human skin fibroblast aggregation[1] which is the unit aggregation area(A/N) (Fig. 1). The abnormal cellular aggregation was observed not only in Alzheimer’s disease [1] but also in other diseases such as Duchenne muscular dystrophy [2], or in Down syndrome [3].

The second idea is the simple majority-rule which is straightforward and means that we can’t diagnose an Alzheimer’s disease case unless we have several bio-markers and a majority agreement, in our case two out of three bio-markers. This idea stems in the complexity of the Alzheimer’s disease which is to often overlooked. In other words we will need to match a complex disease such as Alzheimer’s disease with an equally complex measure, therefore the need for cross-validation. This was our strategy in the past [1] when we looked at fractal and lacunarity analyses for skin fibroblast networks, cell migration, and cell aggregation (A/N) as a bio-marker profile rather than considering each single bio-marker. The same strategy is emphasized in this current paper and we believe this is also a strategy for the future.

The third idea is the quantification of cross-validation. The bio-markers presented in this paper measure different imbalances in human skin fibroblasts, extracellular-signal-regulated kinases (ERK1/2) [4-7], cellular network dynamics [1], or cellular aggregation[1], therefore the need to bring these measures into a common space. This process is called normalization and makes the bio-marker values between zero an one. When all three bio-markers are between zero and one one can compare them in terms of how distant are one from the other, how they synchronize, or how they perform on the case-by-case basis.

The forth idea presented in the paper is the quantification of the cross-validation which uses the Cut-Off not only as a boundary between Alzheimer’s diseases cases and age-matched control cases but also as as normalizing factor. We normalized the values of the bio-markers (x) by subtracting the cut-off value then dividing by the absolute value of the difference, (x-Cutoff)/|x-Cutoff|. Normalized in this way the bio-markers output can only have three values -1, unceartain (U) when x=Cutoff, and 1. In this way the comparison between bio-markers is easy and so is the cross-validation because it comes in the end to a count of the -1, U, or +1.

Last but not least we introduced the concept of hyper-validated cases in the field of Alzheimer’s disease diagnostic. By hyper-validated cases we mean autopsy and/or genetically confirmed Alzheimer’s Disease or non-Alzheimer’s disease demented patients (Non-ADD), and non-demented age-matched controls. In our paper 82% of the cases are hyper-validated cases.


CF fig1

Fig 1. Increased aggregates in Alzheimer’s disease skin fibroblasts. Area per number of aggregates(A/N). Examples of Alzheimer’s disease (AD) fibroblasts aggregates (B), and age-matched controls(AC) (A) at 48 h. Scale bar is 10 µm.



We would like to thank Dr. Camilla Forssten and Dr. Gerhard Nebe-von-Caron from Alere Inc. for their ongoing valuable discussions and Alere Diagnostic for their alliance support. Finally, we would appreciate Dr. Shirley Neitch from Marshall University, Huntington, WV, for her oversight and the clinical diagnosis for the 29 fresh samples presented in this study.



[1] Chirila F.V., Khan T. K., and Alkon D. L. “Spatiotemporal Complexity of Fibroblast Networks Screens for Alzheimer’s Disease”, J Alzheomer’s Disease 33, 165-176 (2013).

[2] Hillier J., Jones G. E., Statham H. E., Witkowski J. A., and Dubowitz V. “Cell surface abnormality in clones of skin fibroblasts from a carrier of Duchenne muscular dystrophy”, J Med Genetics, 22, 100-103 (1985).

[3] Wright T. C., Orkin R. W., Destrempes M., and Kurnit D. M. ”Increased adhesiveness of Down syndrome fetal fibroblasts in vitro”, Proc. Natl. Acad. Sci. USA 81, 2426-2430 (1984).

[4] Zhao W. Q., Ravindranath L., Mohamed A. S., Zohar O., Chen C. H., Lyketsos C. G., Etcheberrigaray R., Alkon D. L. “MAP kinase signaling cascade dysfunction specific to Alzheimer’s disease in fibroblasts”, Neurobiol Dis 11, 166-183 (2002).

[5] Khan T. K., Alkon D. L. “An internally controlled peripheral biomarker for Alzheimer’s disease: Erk1 and Erk2 responses to the inflammatory signal bradykinin”, Proc Natl Acad Sci U S A 103(35), 13203-7 (2006).

[6] Khan T. K., Alkon D. L. “Early diagnostic accuracy and pathophysiologic relevance of an autopsy-confirmed Alzheimer’s disease peripheral biomarker”, Neurobiol Aging 31(6), 889-900 (2008).

[7] Khan T. K., Nelson T. J., Verma V. A., Wender P. A., Alkon D. L. “A cellular model of Alzheimer’s disease therapeutic efficacy: PKC activation reverses Abeta-induced biomarker abnormality on cultured fibroblasts”, Neurobiol Dis. 34(2), 332-9 (2009).



Multiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier SchönmannMultiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier Schönmann