CLIN. CHEM. 27/10,
1712-1716
(1981)
Many Hitherto Unknown Peptides Are Principal Constituents of Uremic
“Middle Molecules”
J. Menyh#{226}rt
and J. Gr#{244}f
The aim of the present investigation was to collect information on the molecular composition of uremic “middle
molecules,” the 500-5000 molecular-mass serum constituents assumed to be involved in the molecular etiology
of uremic intoxication. For this purpose, three fractions
of serum containing such molecules were separated by
cation-exchange column chromatography. These fractions
absorb at 240 nm and are characteristically
increased in
chronic uremia. By one-dimensional paper chromatography of these fractions ft was demonstrated that each couki
be resolved into at least seven (altogether, 21) ninhydrinpositive subfractions. These, like the original fractions,
yielded amino acids on acid hydrolysis. The qualitative
amino acid composition of the subfractions differed substantially, both from each other and from that of 31 known
peptides with which they were compared. We conclude
that hitherto unknown peptides with only limited diffusibility
through hemodialysis membranes constitute the main bulk
of uremic middle molecules.
AdditIonal Keyphrases:
chromatography, paper
chromatography,
cation-exchange
uremia “toxins”
solved
by one-step
cation-exchange
chromatography
into
substantially more components than by gel chromatography
on SG-25, a method excellently suited for bulk separation of
serum MMs (12-14). In addition, it was also demonstrated
that, in uremic intoxication, the solute content of the MMcontaining fractions of serum that could be separated on
Dowex resin (D-fractions), with absorbance at 240 nm, in-
creased characteristically-and
often excessively-in
only
three (D-1, D-X1 and D-X3) of the six D-fractions
(11-16).
It was our aim in the present investigation to get some insight into the molecular composition of serum MMs included
in the uremic D-fractions mentioned above. We show that
peptides of low molecular mass are their principal building
blocks. The peptides were assessedby paperchromatography
and their minimum number in individual D-fractions was
determined. The isolated peptides were characterized by their
qualitative amino acid composition as assessed by paper
chromatography after acid hydrolysis, and by their minimum
molecular mass as estimated on the basis of their qualitative
amino acid composition.
The diffusibility of these uremic peptides through artificial
hemodialysis membranes was also evaluated.
.
Materials and Methods
In 1972, Babb et al. (1) extended the already long list of
putative uremic toxins by adding a new group of serum components to the list. Because at that time these authors had no
idea what the nature of these serum components
might be
chemically,
they termed them “middle
sized” or simply
“middle molecules” (MMs), to emphasize the assumed molecular mass of MMs, which, on the basis of their diffusibility
through hemodialysis
membranes, was thought to range between 300 and 5000 daltons (1). The idea that some of the
clinical symptoms
accompanying
chronic uremia may be
causally related to MMs has been incorporated
into the
“middle-molecule
hypothesis”
(1), which has gained popularity among nephrologists during the past few years. A causal
relationship
between MMs and uremia-related
neuropathy
(3), anemia (4,5), pericarditis (6), increased sensitivity toward
infections (7), and disorders of carbohydrate
metabolism
(8)
and of the immune defense mechanism
(9, 10) have been
postulated. However, there is a striking contradiction between,
on the one hand, the view widely held among clinicians, who
attribute clinical (toxic) significance to the MMs, and, on the
other, our present lack of knowledge in regard to the biology,
pharmacology,
metabolism,
origin, and especially the molecular composition
of this particular
group of substances.
During the last few years, it gradually
became clear that
chemical and biological characterization
of MMs was a prerequisite to any meaningful conclusions on their clinical significance.
Previously
it was shown
(11)
that serum MMs could be re-
United Research Organization of the Hungarian Academy of Sciences and the Semmeiweis University
School of Medicine; Department of Clinical Biochemistry,
Budapest, Hungary.
Address correspondence
to J.M. at: MTA-SOTE
EKSZ, Ull#{246}i
ut
78/a, Budapest H-1082, Hungary.
Received April 1, 1981; accepted June 18, 1981.
1712
CLINICAL CHEMISTRY, Vol. 27, No. 10, 1981
Collection of serum samples: We sampled serum from 40
uremic patients with a creatinine
clearance
of <1, both before
(Ub) and immediately
after (U5) the hemodialysis
treatment
(thrice a week; Cuprophane PT 150 membrane, Kiil dialyzer,
8-h duration).The samples were stored as separate Ub and Ua
pools at -20 #{176}C.
Column chromatography:
Crude 30-mL serum specimens
were chromatographed
on a column of Dowex SOW X 12 cation-exchange
resin as previously
described (14, 15). After
desalting(SG-10, batch technique),the D-fractionswere lyophilized and stored in ampuls, under nitrogen, at -20 #{176}C.
Paper chromatography:
The lyophilized
D-fractionsof
both Ub and U,, were dissolved in0.3mL of distilled water (the
solute content in 0.1 mL of this solution corresponded to that
in 10 mL of the original native serum pools). We applied 50-1zL
samples in duplicate on a 20 X 20 cm sheet of a Whatman
3MM paper.One-dimensionaldouble-ascension
paper chro-
matography was performed for 2 X 4 h at room temperature
inthe system butanol/water/glacial aceticacid(4/1.7/1by vol).
At the end of the run, the paper was cut into longitudinal
strips.
Alternate stripswere developed with ninhydrin reagent
(0.4g of ninhydrin in 100 mL of 96% ethanol), then airdried.
Acid hydrolysis:
The remaining strips with the undeveloped
samples were cut transverselyat heights corresponding to
those of theninhydrin-positive
spots on the developed strips
and then cut into zones. The individual pieces were placed in
1.0 mL of water and gentlyagitatedforthe next 24 h. The
supernatant fluid and 1.5 mL of distilledwater, added in
0.5-mL portions forrepeated washing of the cut paper strips,
were pooled, transferred into ampuls, and lyophilized. Into
these ampuls,
and into those containing
the lyophilized
D(seeabove),we pipetted0.1mL of6 mol/L
fractions
ampuls
hydrolyzed
HC1. The
N2, sealed, and their contents were
at 100 #{176}C
for 24 h.
were gassed
with
Qualitative
amino acid analysis:
After the HC1 had been
removed by repeated evaporation under reduced pressure and
re-solution in water, the final residue was redissolved in 100
zL of distilled water. This and a similar volume of a standard
amino acid mixture containing 16 known amino acids [cysteine
(Cys), lysine (Lys), histidine (His), arginine (Arg), glycine
(Gly), serine (Ser), asparagine (Asn), threonine (Thr), glutsmine (Gin), alanine (Ma), proline (Pro), tyrosine (Tyr), methionine (Met), valine (Val), phenylalanine (Phe), leucine
(Leu)]were appliedto Whatman 3MM paper under conditions identical with thosealreadydescribed.
To enhance the
detectability of those amino acids present in the lowest concentration, we occasionally simultaneously applied three
different volumes of samples with solute content corresponding to that in 5, 10, and 15 mL of the original serum.
0
GO
F
Gil
GQ
F
E
c
ocD
U
cg0
Results
Paper Chromatography
C
B
__________
front
B
B
of the D-Fractions
Part A of Figure 1 shows that Ub-derived D-1 fractioncould
be resolvedby paper chromatography intoseven ninhydrinpositivesubfractions.Elongated profilesof subfractionsB,
D, and E suggest their heterogeneous composition. The
smaller number ofninhydrin-positive
spotson chromatograms
of the U,,-derivedD-1 fractionindicatedthat two (E and F)
of the seven subfractionsin Ut,-derivedD-1 fractioncould no
longer be detected afterroutine hemodialysis treatment.
Part B of Figure 1 demonstratesthatthe Ub-derived D-X1
fraction
could similarly
be resolved
into seven ninhydrinpositive subfractions.The heterogeneityof subfractionC can
be assumed. Chromatograms of the Ua-derived D-X1 fraction
indicate that five (A, B, D, E, and F) of the seven Ub-derived
subfractions could be removed by routine hemodialysis
treatment.
Like the D-1 and D-X1 fractions,the Ub-derived D-X3
fractionalsocould be resolvedintoseven ninhydrin-positive
subfractions(PartC of Figure 1).Heterogeneityofsubfraction
G ishighly probable. After hemodialysis subfractionsA and
C, and perhaps subfractionE, were no longerdetectable,and
subfractionsB, F, and G were diminished. In contrast,hemodialysis was practically
ineffective
in removing subfraction
B. The assumed heterogeneityof subfractionD in Ub-derived
D-X3 fractionwas furthersubstantiatedby its resolutioninto
two separate subfractionsin samples collectedafterhemodialysis.
Qualitative Amino-Acid Composition of Acid
Hydrolysates of the D-Fractions and of Their PaperChromatographic
Subfractions
All D-fractions characteristic
of chronic uremia (D-1, D-X1,
and D-X3) and all of their paper-chromatographically
resolved
subfractions
yielded amino acids upon acid hydrolysis, demonstrating
the peptide character of these building
blocks
(D-peptides).
Table 1 summarizes the qualitative amino-acid composition
of the Ub-derived
D-1, D-X1, and D-X3 fractions as well as
that of their paper-chromatographic
subfractions
(A-G). In
the D-1 fraction, 14 of the 16 standard amino acids could be
identified;
His and Val were absent. Besides the identified
amino acids, one unidentified ninhydrin-positive
component
with a low Rf value was also detected in this fraction. The
number
of identified
amino acids in the subfractions
derived
from the D-1 fraction
varied between five (subfraction
A and
B) and nine (subfraction
G). The unidentified
ninhydrinpositive
component
detected
in the D-1 fraction was found
in all of the subfractions.
The relative frequency
of occurrence
of the identified
amino acids varied considerably
from subfraction
to subfraction:
Cys and Ser were found in six; Gly and
A80
A
start
Ub
Ua
1
UbUa
xl
Ub
Ua
x3
Fig. 1. Paper-dwomatographicprofile (developed with ninhy&in)
of the main uremia-related serum fraction 1, X1, and X3 obtained
with Dowex cation-exchange chromatography
Ub.
serum obtained before, and Ua,serum obtained after hemodlalysis of the
patient
Asn in four; Arg, Pro, Ala, and Thr in three; Lys and Phe in
two subfractions. Leu was present in one subfraction only. In
subfraction A, E, and G, it was not clear whether Thr or Gln
was the amino acid residue. Calculations based on the molecular mass of the amino acid residues and on the assumption
that there was only one amino acid residue for each amino acid
detected in the subfractions indicated that the lowest molecular mass of peptides in the D-1 fraction ranged between
529 or 556 (subfraction A) and 1046 or 1043 (subfraction G)
daltons, depending on whether Thr or Gin was calculated as
an amino acid residue (reaction water was deducted from the
summed molecular mass of the amino acid residues).
Table 1 also shows the qualitative amino acid composition
of the Ub-derived D-X1 fraction and that of its subfractions.
Of the 16 standard amino acids,12 could be identified in this
fraction. Ala, Pro, Val, and Leu were missing. Beside those
amino acids identified, two unidentified ninhydrin-positive
components, both with a low Rf value, were found in this
fraction. The number of identified amino acids present in
subfractions derived from the D-X1 fraction varied between
two (subfraction G) and 10 (subfraction E). Both of the unidentified ninhydrin-positive components were present in one
(A); only one was present in four (B, C, D, and E) subfractions.
The relative frequency of occurrence of the identified amino
acid residues in the various subfractions was as follows: Asn
and Gly were present in four; Lys, Ser, Tyr, and Phe in three;
and His, Thr, and Gln in two subfractions. Cys, Arg, and Met
were detected in a single subfraction only. In subfraction B
and D it remained undecided whether Gly or Ser (subfraction
B), His or Arg (subfraction D) were the amino acid residues.
Because of insufficient sample, amino acid analysis of subfraction F had to be omitted. Based on similar assumptions
and calculation as previously mentioned, the minimum molecular masses of the various peptides in the D-X1 fraction
were estimated to lie between 189 (subfraction
G) and 1182
(subfraction C) daltons, respectively.
The qualitative amino acid composition of the Ub-derived
CLINICAL CHEMISTRY, Vol. 27, No. 10, 1981
1713
CO
nIl
l+I+l+++l+l
II
l++#{176}
1
(0
iO
.+I
I+l++l++l
‘5
II
l+l++
I1+11+l+1+lIIIll
in
l’
II.
11+11
l+l++I
I
I+I++FC’)
I+I+l++I
.
11111
C
(5
‘r
I
0’-
ho
1111
I
_
I +
I
I
I + + + +
I
I
I
+l++l+++++++(
I
I
lir-
in
C’)
llin
0
CO
.
CO
l+l++l
I
.
g
.5,
0)
CO
lllIll-I-l-4-IllllIlIlc.
o2
013
#{149}g
a
‘-
C(5
ILlllllllllllllllllll
I
.
CO
.
Wl+++l
I+++++l
1+1
l+1o
1+1
l+I
0
a
U)U)
0
-
(5 0x
-(5
-
ol+I
l+++l+l
0J
CO
l++I+1a
l+l
I+l()$l
‘-
11111111
lc1’t
#{149}0
4++l+l
01
01
01
01
:2
a
C’)
I+lIlIllllllIc’J
0
01
tS’-
I
.C0#{149}
-
+1
I ++
I +
I+I+l++()()++l+l
I
W+l+l
II
I
I
I
CO
C’)
,-
‘-
.5
I+a
l++I++l
l+l+ICO
l+++I
I+l+l-
0)
F-
a+l+l
+1+1
l+++I+I
1+1111
1CO
l+++l+I
1+1111
ICO
11+111
CO
l’#{176}0
+I+Il
l+l++I
+1+11
I ++
l++ll
‘-
Co
1+11
CO 0
(5
F-
1-
0)
0)
5
0)
‘
CO
.5
0
0)
0)
a
CO
in”,
I
I
a
.
F-
E
a
o
2-
o(5.
(0
.
llu,01in
I ++
IS)
‘
IS)
I ++++++++++
0) 01 15) CO 0)
0inCODCOF-0
o
0)
CO
5UC-
e
._
w
I-.
D-X3 fraction and that of its subfractions are seen in Table
1. In the D-X3 fraction, 12 of the 16 standard amino acids
could be detected. Those lacking were His, Pro, Tyr, and Val.
One unidentified ninhydrin-positive
component with a low
Rf value was found in this fraction. The number of identified
amino acids in the individual subfractions varied between four
(subfraction
E) and nine (subfraction
G). The unidentified
ninhydrin-positive
component
in the D-X3 fraction was found
1714
CLINICAL CHEMISTRY,
Vol. 27. No. 10. 1981
in three of the seven subfractions.
The relative frequency of
occurrence
of the identified
amino acids in the various subfractions was as follows: Thr was present in six; Lys, Ser,and
Asn in five; Arg and GIn in four; Cys, Gly, Met, Phe, and Leu
in three subfractions;
and Ala in only one subfraction.
Calculated as mentioned
above, the lowest molecular mass of the
peptides in the D-X3 fraction varied between 445 (subfraction
E) and 1066 (subfraction G) daltons.
Dscusslon
Thrice-repeated
similar experiments provided results that
were essentially identical. Thus a remarkably large number
of peptides of low molecular mass evidently are the principal
building blocks of the D-fractions derived from uremic sera.
This conclusion
was based on observations that each of the
three uremic D-fractions could be resolved by one-dimensional paper chromatography into at least seven (altogether
21 different) ninhydrin-positive
spots, which yielded amino
acids upon acid hydrolysis. We are fully aware that onedimensional paper chromatography is far from being the most
adequate method for purification
to homogenicity of peptides
from a solution of heterogeneous composition. Thus, heterogeneity of at least some of the subfractions discussed in this
paper cannot be excluded. However, it is worth mentioning
that, with the sole exception of the D-1 derived subfraction
C and D, the qualitative amino acid composition of the rest
of the subfractions consistently differed from each other. It
is also noteworthy that most of the subfractions could be
categorized in two separate groups on the basis of hemodialysis: one eliminated and the other completely unaltered by
hemodialysis. Both findings indicate that these subfractions
may represent distinct molecular entities with closely related
physicochemical properties. However, it should be borne in
mind that the principal message stressed in this paper remains
valid irrespective of the purity of the subfractions. The message is that, with even such a simple, not to say primitive,
technique as one-dimensional paper chromatography, uremic
D-fractions could be resolved into at least 21 distinct peptides
(or groups of peptides) with qualitatively different amino acid
composition. We stress that21 is the minimum number of
peptides present in these subfractions.
Besides the two to nine amino acid residues identified, one
or two ninhydrin-positive
unidentified substances could also
be detected in each subfraction. Their conspicuously low Rf
values (lower than those of the basic amino acids) suggest an
unusually strong basic character of these components. They
may represent polyamines; spermidine was shown to be attached to some of the uremic peptides in a currently unknown
manner (13).
The minimum molecular masses calculated for the uremic
D-peptides ranged between 189 and 1066 daltons. Circumstantial evidence indicates, however, that the actual molecular
mass of the uremic D-peptides may not differ substantially
from that we estimated. For example, molecular masses of the
D-fractions as also estimated on a calibrated analytical SG-25
column were similar to those calculated as mentioned above
(unpublished data). Furthermore, results obtained by ultrafiltration of the D-fractions through Amicon YM and UM 05
membranes also indicated molecular masses ranging between
500 and 5000 daltons (unpublished data). Finally, it is worth
mentioning that the molecular mass of the two uremic peptides of Abiko et al. (22) also falls into a similar range (see
below). Taking all these facts together, it seems reasonable
to assume that the molecular mass of theuremic D-peptides
fallswithin the range of that of the MMs. On thisbasis,Dpeptides should be regarded as constituents of serum MMs.
The molecular mass of a significant number of the known
serum peptides also falls within the range of serum MMs.
Thus, by definition, they must be counted among the MMs.
The question arises as to the relationship between the known
serum peptides normally present and the uremic MMs. No
doubt, the concentrations of several known peptide hormones
increase in chronic uremia (17). It was even speculated that
such a situation may eventually lead to serious consequences
and contribute to the development of some of the clinical
syndromes in uremic intoxication. For instance, Bricker’s
“trade-off hypothesis” is based on such an assumption (18).
On the other hand, several direct and circumstantial pieces
of evidence suggest that abnormal peptides also may additionally appear and accumulate in uremic body fluids. The
best evidence of this kind was presented by Abiko et al. (22),
whose work led not only to the identification of two uremiarelated and hitherto unknown molecules, a tryptophan-contaming pentapeptide (19) and a heptapeptide (20), but also
to the formulation of one of the possible molecular mechanisms participating in the bioproduction of abnormal peptides. These two peptides were shown to be split products of
fibrinogen (pentapeptide) and globulin (heptapeptide) molecules. Conceivably, uremic intoxication may create favorable
conditions for splitting protein molecules at ordinarily forbidden sites. Under these conditions, abnormal peptide sequences may appear and accumulate in the body fluids, with
thehandling ofwhich the organism is unfamiliar and against
which, iftoxic,
itis unprotected. This kind of metabolic manipulation of the protein molecules may represent a molecular
mechanism of universal validity that underlies not only the
molecular etiology of uremic intoxication but also that of a
much broader spectrum of endogenous intoxications.
There are two bits of circumstantial evidence that make it
unlikely that D-peptides would be identical with any of the
known serum peptides. One is that the qualitative amino acid
composition of the D-peptides was found to be different from
that of 31 known regulatory peptides with molecular mass of
about 5000 daltons or less [ACTH, angiotensins, bombesin,
caerulein, calcitonin, carnosine, CCK-octapeptide, C-peptide,
CRH, delta-sleep inducing peptide, endorphins, enkephalins,
gastrin I (human), glucagon, insulin, LHRH, liver cell growth
factor, MIF, motilin, MSHs, neurotensin, oxytocin, pepatatin,
secretin, somatostatin, Substance P, thymus hormone, TRH,
tuftain, vasopressin, VIP]. However, the uncertainties existing
in regard to the purity of the D-peptides make this argument
of questionable value. On the other hand, the surprisingly high
serum concentration of D-peptides, which is simply incomparable to that of the known regulatory peptides, is a strong
argument in favor of their dissimilarity to the known serum
peptides. While the concentration of the latter in serum ranges
between 10b0 and iO molfL, D-peptides are represented
in about 10-6 molfL concentration in the serum.
We conclude thatMMs, the concentrations of which are
significantly
increased in serum of chronic uremics, are
principally made up of a remarkably large number of peptides.
Routine hemodialysis seems to completely remove some of
these peptides, while others are only partly removed or left
completely unaltered by such treatment.
Several authors reported in the past on various biological
effects of different serum fractions composed entirely or partly
of MM ingredients. Much of these effects simulated some of
the clinical symptoms accompanying uremic toxemia (3-9,
15,16,20-22). Data citedand presented in this paper provide
a good reason to regard D-peptides as the putative molecular
carriers of at least some of these symptoms. Future research
should reveal the clinical significance of the individual Dpeptides in eliciting and (or)maintaining various clinical
symptoms in uremic intoxication.
We are grateful to Mrs. A. Gr#{243}f
and E. Korompai fortheirexpert
technical assistance and to Mrs. P. Pataky fortypingthe manuscript.
References
1. Babb, A. L., Farrell, P. C., Uvelli, D. A., and Scribner, B. M., He-
modialyser evaluation by examination
of solute
molecular spectra.
Am. Soc. Artif. Intern. Organs 17,98-105
(1972).
2. Millora, A.B., Woodruff, M. W.,and Kiley, J. E., Comparison of
Trans.
CLINICAL CHEMISTRY,
Vol. 27, No. 10, 1981
1715
higher blood flowhemodialysis with lower blood flow in light of the
square meter-hour
hypothesis. Ibid. 18,85-89 (1972).
3. Funck-Brentano,
J. L.,Man, N. K.,Sausse, A.,et al., Characterization of a 1100-1300 MW uremic neurotoxin.
Ibid. 22, 163-167
(1976).
4. Goubeaud, G.,Leber,H. W., Schott,H. H.,and SchOtterle, G.,
Middle
molecules
and hemoglobin
synthesis.
Proc. Eur. Dial.
Transplant Assoc. 13, 371-376 (1977).
5. Goubeaud, G., Leber, H. W., and Schutterle, G., Trennmethoden
und klinische
Relevanz von sogenannten
Mittehnolekulen
in der
UrAmie. In 2. Donausymposium
f#{252}r
Nephrologie, B. Watschinger,
Ed.,Verlag Carl Bindernagel, Friedberg-Hessen, 1978,pp 99-104.
6. Bergstrom, J., Uremic toxicity. Proc. Eur. Dial. Transplant Assoc.
12,579-588 (1975).
7. Traeger, J., Touraine,
J. L.,Revillard, J. P., et al., Role of
methylguanidine
and middle moleculesin the immune deficiency
secondaryto uremia.In Proc. 6th mt. Congr. Nephrology. S. Gbvanetti,
V. Bonomini,and G. D. Amico,Eds., S. Karger,Basel, 1975,
pp 584-590.
8. Dzurik,R.,Pathophysiologie
ofthecarbohydratemetabolism in
renalinsufficiency.
Czech. Med. 20, 179-188 (1974).
9. Touraine, J. L., Navarro, J., Lorre,C., and Traeger, J., Inhibitory
effect
of medium sized molecules from patients with renal failure on
lymphocyte stimulation by phytohemagglutinin. Biomedicine
23,
Acta 54, 47-54 (1974).
13. Lutz, W., Studies on formation
of complexes
of
peptides isolated from the plasmaof uremic patients.
16, 159-170
insulin with basic
Acta Med. Pol.
(1975).
14. Chany, T. M. S., Michelsen, M., Coffey, J. F., and Stark, A., Serum
middle molecule levels in uremia during long term intermittent hemoperfusion with the ACAC/coated charcoal/microcapsule
artificial
kidney. Trans. Am. Soc. Artif. Intern. Organs 20,364-371
(1974).
15. Gr#{243}f,
J.,Menyh#{225}rt,
J.,Marcsek, Z., andBabics,A., Vergleichende
Untersuchung der Polypeptidfractionen
in von gesunden und ur#{227}mischen Personenstammenden Seren.Exp. Med. Sci. (Budapest)
26,540-549 (1974).
16.Menyh#{225}rt,
J.,and Gr#{243}f,
J., Composition and diffusibility
of
middle molecularweightsubstances(MMS) preparedfrom seraof
healthy and uremic individuals. J. Mol. Med. 2,371-380 (1977).
17. Migone, L., Dall’Aglio, P., and Buzio, C., Middle molecules in
uremic serum, urineand dialysis
fluid.Clin. Nephrol. 3, 82.93
(1975).
18. Bricker, N. S., On the pathogenesisof the uremic state: An exposition of the “trade-off hypothesis.” N. Engl. J. Med. 286, 10931099 (1972).
19. Abiko, T., Kumikava, M., Higuchi, H., and Sekino, H., Identification and synthesis of a heptapeptide in uremic fluid. Biochem.
B.
Biophys. Res. Commun. 84, 184-194 (1978).
20. Abiko, T.,Onodea, T., and Sekino, H., Isolation, structure and
biological activity
ofthetryptophane-containing
pentapeptidefrom
uremic fluid. Ibid. 89,813-821(1979).
21. Lutz, W., Chemical composition and rate of passage across
semipermeablemembranes of basic peptidesisolated
from peritoneal
dialysis fluid
from patients
withchronicrenal failure. Acta Med. Pol.
Watsehinger, Ed., Verlag Carl Bindernagel, Friedberg-Hessen, 1979,
pp 123-134.
12. Hanicki, Z., Sarnecka-Keller, M., Klein, A.,and Slizowska, K.,
Middle-sized ninhydrin-positive moleculesinuremic patients treated
by repeated hemodialysis.
I. Preliminary
characteristics.
Clin. Chim.
17, 137-147 (1976).
22. Abiko, T., Onodea, T. J., Sekino, H., and Higuchi, H., Effect of
synthetic serum thymic factor fragments on the inhibition
of sheep
erythrocyte rosette formation by uremic toxin. Chem. Pharm. Bull.
28, 667-669 (1980).
180-184
(1975).
10. Cichocki, T., Hanicki, Z., Klein, A.,etal., Influence of middlemolecular-weight solutes from dialysate on the migration rate of
leukocytes.
Kidney
Int. 17, 231-236
(1980).
11. Gr#{243}f,
J.,and Menyh#{225}rt,
J.,Preparative
fractionationaton
middle molecules.
In 3. Donausymposium
f#{252}r
Nephrologie,
1716
CLINICAL CHEMISTRY,
Vol. 27, No. 10, 1981
of